KIARA_METADATA
¶
Functions¶
DBG(*objects, *, sep=' ', end='\n', file=None, flush=False)
¶
Source code in kiara/__init__.py
def DBG(
*objects: typing.Any,
sep: str = " ",
end: str = "\n",
file: typing.Optional[typing.IO[str]] = None,
flush: bool = False,
):
objs = (
["[green]----------------------------------------------[/green]"]
+ list(objects)
+ ["[green]----------------------------------------------[/green]"]
)
dbg(*objs, sep=sep, end=end, file=file, flush=flush)
dbg(*objects, *, sep=' ', end='\n', file=None, flush=False)
¶
Source code in kiara/__init__.py
def dbg(
*objects: typing.Any,
sep: str = " ",
end: str = "\n",
file: typing.Optional[typing.IO[str]] = None,
flush: bool = False,
):
for obj in objects:
try:
rich_print(obj, sep=sep, end=end, file=file, flush=flush)
except Exception:
rich_print(
f"[green]{obj}[/green]", sep=sep, end=end, file=file, flush=flush
)
get_version()
¶
Return the current version of Kiara.
Source code in kiara/__init__.py
def get_version() -> str:
"""Return the current version of *Kiara*."""
from pkg_resources import DistributionNotFound, get_distribution
try:
# Change here if project is renamed and does not equal the package name
dist_name = __name__
__version__ = get_distribution(dist_name).version
except DistributionNotFound:
try:
version_file = os.path.join(os.path.dirname(__file__), "version.txt")
if os.path.exists(version_file):
with open(version_file, encoding="utf-8") as vf:
__version__ = vf.read()
else:
__version__ = "unknown"
except (Exception):
pass
if __version__ is None:
__version__ = "unknown"
return __version__
Modules¶
context
special
¶
logger
¶
Classes¶
Kiara
¶
The core context of a kiara session.
The Kiara object holds all information related to the current environment the user does works in. This includes:
- available modules, operations & pipelines
- available value data_types
- available metadata schemas
- available data items
- available controller and processor data_types
- misc. configuration options
It's possible to use kiara without ever manually touching the 'Kiara' class, by default all relevant classes and functions
will use a default instance of this class (available via the Kiara.instance() method.
The Kiara class is highly dependent on the Python environment it lives in, because it auto-discovers available sub-classes of its building blocks (modules, value data_types, etc.). So, you can't assume that, for example, a pipeline you create will work the same way (or at all) in a different environment. kiara will always be able to tell you all the details of this environment, though, and it will attach those details to things like data, so there is always a record of how something was created, and in which environment.
Source code in kiara/context/__init__.py
class Kiara(object):
"""The core context of a kiara session.
The `Kiara` object holds all information related to the current environment the user does works in. This includes:
- available modules, operations & pipelines
- available value data_types
- available metadata schemas
- available data items
- available controller and processor data_types
- misc. configuration options
It's possible to use *kiara* without ever manually touching the 'Kiara' class, by default all relevant classes and functions
will use a default instance of this class (available via the `Kiara.instance()` method.
The Kiara class is highly dependent on the Python environment it lives in, because it auto-discovers available sub-classes
of its building blocks (modules, value data_types, etc.). So, you can't assume that, for example, a pipeline you create
will work the same way (or at all) in a different environment. *kiara* will always be able to tell you all the details
of this environment, though, and it will attach those details to things like data, so there is always a record of
how something was created, and in which environment.
"""
_instance = None
@classmethod
def instance(cls) -> "Kiara":
"""The default *kiara* context. In most cases, it's recommended you create and manage your own, though."""
if cls._instance is None:
cls._instance = Kiara()
return cls._instance
@classmethod
def create_in_path(cls, path: Union[str, Path]):
config = KiaraConfig(base_path=path)
return cls.create(config=config)
@classmethod
def create(
cls,
config: Union[None, str, Path, KiaraConfig] = None,
context_name: Optional[str] = None,
**extra_context_args: Any,
):
if config is None:
config = KiaraConfig()
elif isinstance(config, str):
if os.path.isfile(config):
config_data = get_data_from_file(config)
if not isinstance(config_data, Mapping):
raise ValueError(
f"Invalid config file format, can't parse file: {config}"
)
config = KiaraConfig(**config_data)
else:
raise Exception(
f"Can't read kiara config from file, file does not exist: {config}"
)
elif isinstance(config, Path):
if config.is_file():
config_data = get_data_from_file(config)
if not isinstance(config_data, Mapping):
raise ValueError(
f"Invalid config file format, can't parse file: {config}"
)
config = KiaraConfig(**config_data)
else:
raise Exception(
f"Can't read kiara config from file, file does not exist: {config.as_posix()}"
)
elif isinstance(config, Mapping):
config = KiaraConfig(**config)
elif not isinstance(config, KiaraConfig):
raise Exception(
f"Can't create kiara context instance, invalid config type: {type(config)}"
)
context = config.create_context(context_name=context_name, **extra_context_args)
return context
def __init__(self, config: Optional[KiaraContextConfig] = None):
if not config:
kc = KiaraCurrentContextConfig()
config = kc.get_context()
self._id: uuid.UUID = ID_REGISTRY.generate(
id=uuid.UUID(config.context_id), obj=self
)
ID_REGISTRY.update_metadata(self._id, kiara_id=self._id)
self._config: KiaraContextConfig = config
if is_debug():
echo = True
else:
echo = False
self._engine: Engine = create_engine(
self._config.db_url,
echo=echo,
future=True,
json_serializer=orm_json_serialize,
json_deserializer=orm_json_deserialize,
)
# self._run_alembic_migrations()
# self._envs: Optional[Mapping[str, EnvironmentOrm]] = None
self._event_registry: EventRegistry = EventRegistry(kiara=self)
self._type_registry: TypeRegistry = TypeRegistry(self)
self._data_registry: DataRegistry = DataRegistry(kiara=self)
self._job_registry: JobRegistry = JobRegistry(kiara=self)
self._module_registry: ModuleRegistry = ModuleRegistry()
self._operation_registry: OperationRegistry = OperationRegistry(kiara=self)
self._alias_registry: AliasRegistry = AliasRegistry(kiara=self)
self._destiny_registry: DestinyRegistry = DestinyRegistry(kiara=self)
self._env_mgmt: Optional[EnvironmentRegistry] = None
metadata_augmenter = CreateMetadataDestinies(kiara=self)
self._event_registry.add_listener(
metadata_augmenter, *metadata_augmenter.supported_event_types()
)
self._context_info: Optional[KiaraContextInfo] = None
# initialize stores
self._archive_types = find_all_archive_types()
self._archives: Dict[str, KiaraArchive] = {}
for archive_alias, archive in self._config.archives.items():
archive_cls = self._archive_types.get(archive.archive_type, None)
if archive_cls is None:
raise Exception(
f"Can't create context: no archive type '{archive.archive_type}' available. Available types: {', '.join(self._archive_types.keys())}"
)
config_cls = archive_cls._config_cls
archive_config = config_cls(**archive.config)
archive_obj = archive_cls(archive_id=archive.archive_uuid, config=archive_config) # type: ignore
for supported_type in archive_obj.supported_item_types():
if supported_type == "data":
self.data_registry.register_data_archive(
archive_obj, alias=archive_alias # type: ignore
)
if supported_type == "job_record":
self.job_registry.register_job_archive(archive_obj, alias=archive_alias) # type: ignore
if supported_type == "alias":
self.alias_registry.register_archive(archive_obj, alias=archive_alias) # type: ignore
def _run_alembic_migrations(self):
script_location = os.path.abspath(KIARA_DB_MIGRATIONS_FOLDER)
dsn = self._config.db_url
log_message("running migration script", script=script_location, db_url=dsn)
from alembic.config import Config
alembic_cfg = Config(KIARA_DB_MIGRATIONS_CONFIG)
alembic_cfg.set_main_option("script_location", script_location)
alembic_cfg.set_main_option("sqlalchemy.url", dsn)
command.upgrade(alembic_cfg, "head")
@property
def id(self) -> uuid.UUID:
return self._id
@property
def context_config(self) -> KiaraContextConfig:
return self._config
@property
def context_info(self) -> "KiaraContextInfo":
if self._context_info is None:
self._context_info = KiaraContextInfo.create_from_kiara_instance(kiara=self)
return self._context_info
# ===================================================================================================
# registry accessors
@property
def environment_registry(self) -> EnvironmentRegistry:
if self._env_mgmt is not None:
return self._env_mgmt
self._env_mgmt = EnvironmentRegistry.instance()
return self._env_mgmt
@property
def type_registry(self) -> TypeRegistry:
return self._type_registry
@property
def module_registry(self) -> ModuleRegistry:
return self._module_registry
@property
def alias_registry(self) -> AliasRegistry:
return self._alias_registry
@property
def destiny_registry(self) -> DestinyRegistry:
return self._destiny_registry
@property
def job_registry(self) -> JobRegistry:
return self._job_registry
@property
def operation_registry(self) -> OperationRegistry:
op_registry = self._operation_registry
return op_registry
@property
def data_registry(self) -> DataRegistry:
return self._data_registry
@property
def event_registry(self) -> EventRegistry:
return self._event_registry
# ===================================================================================================
# context specific types & instances
@property
def current_environments(self) -> Mapping[str, RuntimeEnvironment]:
return self.environment_registry.environments
@property
def data_type_classes(self) -> Mapping[str, Type[DataType]]:
return self.type_registry.data_type_classes
@property
def data_type_names(self) -> List[str]:
return self.type_registry.data_type_names
@property
def module_type_classes(self) -> Mapping[str, Type["KiaraModule"]]:
return self._module_registry.module_types
@property
def module_type_names(self) -> Iterable[str]:
return self._module_registry.get_module_type_names()
# ===================================================================================================
# kiara session API methods
def create_manifest(
self, module_or_operation: str, config: Optional[Mapping[str, Any]] = None
) -> Manifest:
if config is None:
config = {}
if module_or_operation in self.module_type_names:
manifest: Manifest = Manifest(
module_type=module_or_operation, module_config=config
)
elif module_or_operation in self.operation_registry.operation_ids:
if config:
raise Exception(
f"Specified run target '{module_or_operation}' is an operation, additional module configuration is not allowed (yet)."
)
manifest = self.operation_registry.get_operation(module_or_operation)
elif os.path.isfile(module_or_operation):
raise NotImplementedError()
else:
raise Exception(
f"Can't assemble operation, invalid operation/module name: {module_or_operation}. Must be registered module or operation name, or file."
)
return manifest
def create_module(self, manifest: Union[Manifest, str]) -> "KiaraModule":
"""Create a [KiaraModule][kiara.module.KiaraModule] object from a module configuration.
Arguments:
manifest: the module configuration
"""
return self._module_registry.create_module(manifest=manifest)
def queue(
self, manifest: Manifest, inputs: Mapping[str, Any], wait: bool = False
) -> uuid.UUID:
"""Queue a job with the specified manifest and inputs.
Arguments:
manifest: the job manifest
inputs: the job inputs
wait: whether to wait for the job to be finished before returning
Returns:
the job id that can be used to look up job status & results
"""
return self.job_registry.execute(manifest=manifest, inputs=inputs, wait=wait)
def process(self, manifest: Manifest, inputs: Mapping[str, Any]) -> ValueMap:
"""Queue a job with the specified manifest and inputs.
Arguments:
manifest: the job manifest
inputs: the job inputs
wait: whether to wait for the job to be finished before returning
Returns
"""
return self.job_registry.execute_and_retrieve(manifest=manifest, inputs=inputs)
def save_values(
self, values: ValueMap, alias_map: Mapping[str, Iterable[str]]
) -> StoreValuesResult:
_values = {}
for field_name in values.field_names:
value = values.get_value_obj(field_name)
_values[field_name] = value
self.data_registry.store_value(value=value, skip_if_exists=True)
stored = {}
for field_name, field_aliases in alias_map.items():
value = _values[field_name]
try:
if field_aliases:
self.alias_registry.register_aliases(value.value_id, *field_aliases)
stored[field_name] = StoreValueResult.construct(
value=value, aliases=sorted(field_aliases), error=None
)
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
stored[field_name] = StoreValueResult.construct(
value=value, aliases=sorted(field_aliases), error=str(e)
)
return StoreValuesResult.construct(__root__=stored)
# def run(self, module_or_operation: str, module_config: Mapping[str, Any] = None):
#
# if isinstance(module_or_operation, str):
# if module_or_operation in self.operation_registry.operation_ids:
#
# operation = self.operation_registry.get_operation(module_or_operation)
# if module_config:
# print(
# f"Specified run target '{module_or_operation}' is an operation, additional module configuration is not allowed."
# )
#
# elif module_or_operation in self.module_type_names:
#
# if module_config is None:
# module_config = {}
# manifest = Manifest(
# module_type=module_or_operation, module_config=module_config
# )
#
# module = self.create_module(manifest=manifest)
# operation = Operation.create_from_module(module)
#
# elif os.path.isfile(module_or_operation):
# raise NotImplementedError()
# # module_name = kiara_obj.register_pipeline_description(
# # module_or_operation, raise_exception=True
# # )
# else:
# merged = list(self.module_type_names)
# merged.extend(self.operation_registry.operation_ids)
# raise NoSuchExecutionTargetException(
# msg=f"Invalid run target name '[i]{module_or_operation}[/i]'. Must be a path to a pipeline file, or one of the available modules/operations.",
# available_targets=sorted(merged),
# )
alias_registry: AliasRegistry
property
readonly
¶context_config: KiaraContextConfig
property
readonly
¶context_info: KiaraContextInfo
property
readonly
¶current_environments: Mapping[str, kiara.models.runtime_environment.RuntimeEnvironment]
property
readonly
¶data_registry: DataRegistry
property
readonly
¶data_type_classes: Mapping[str, Type[kiara.data_types.DataType]]
property
readonly
¶data_type_names: List[str]
property
readonly
¶destiny_registry: DestinyRegistry
property
readonly
¶environment_registry: EnvironmentRegistry
property
readonly
¶event_registry: EventRegistry
property
readonly
¶id: UUID
property
readonly
¶job_registry: JobRegistry
property
readonly
¶module_registry: ModuleRegistry
property
readonly
¶module_type_classes: Mapping[str, Type[KiaraModule]]
property
readonly
¶module_type_names: Iterable[str]
property
readonly
¶operation_registry: OperationRegistry
property
readonly
¶type_registry: TypeRegistry
property
readonly
¶Methods¶
create(config=None, context_name=None, **extra_context_args)
classmethod
¶Source code in kiara/context/__init__.py
@classmethod
def create(
cls,
config: Union[None, str, Path, KiaraConfig] = None,
context_name: Optional[str] = None,
**extra_context_args: Any,
):
if config is None:
config = KiaraConfig()
elif isinstance(config, str):
if os.path.isfile(config):
config_data = get_data_from_file(config)
if not isinstance(config_data, Mapping):
raise ValueError(
f"Invalid config file format, can't parse file: {config}"
)
config = KiaraConfig(**config_data)
else:
raise Exception(
f"Can't read kiara config from file, file does not exist: {config}"
)
elif isinstance(config, Path):
if config.is_file():
config_data = get_data_from_file(config)
if not isinstance(config_data, Mapping):
raise ValueError(
f"Invalid config file format, can't parse file: {config}"
)
config = KiaraConfig(**config_data)
else:
raise Exception(
f"Can't read kiara config from file, file does not exist: {config.as_posix()}"
)
elif isinstance(config, Mapping):
config = KiaraConfig(**config)
elif not isinstance(config, KiaraConfig):
raise Exception(
f"Can't create kiara context instance, invalid config type: {type(config)}"
)
context = config.create_context(context_name=context_name, **extra_context_args)
return context
create_in_path(path)
classmethod
¶Source code in kiara/context/__init__.py
@classmethod
def create_in_path(cls, path: Union[str, Path]):
config = KiaraConfig(base_path=path)
return cls.create(config=config)
create_manifest(self, module_or_operation, config=None)
¶Source code in kiara/context/__init__.py
def create_manifest(
self, module_or_operation: str, config: Optional[Mapping[str, Any]] = None
) -> Manifest:
if config is None:
config = {}
if module_or_operation in self.module_type_names:
manifest: Manifest = Manifest(
module_type=module_or_operation, module_config=config
)
elif module_or_operation in self.operation_registry.operation_ids:
if config:
raise Exception(
f"Specified run target '{module_or_operation}' is an operation, additional module configuration is not allowed (yet)."
)
manifest = self.operation_registry.get_operation(module_or_operation)
elif os.path.isfile(module_or_operation):
raise NotImplementedError()
else:
raise Exception(
f"Can't assemble operation, invalid operation/module name: {module_or_operation}. Must be registered module or operation name, or file."
)
return manifest
create_module(self, manifest)
¶Create a [KiaraModule][kiara.module.KiaraModule] object from a module configuration.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
manifest |
Union[kiara.models.module.manifest.Manifest, str] |
the module configuration |
required |
Source code in kiara/context/__init__.py
def create_module(self, manifest: Union[Manifest, str]) -> "KiaraModule":
"""Create a [KiaraModule][kiara.module.KiaraModule] object from a module configuration.
Arguments:
manifest: the module configuration
"""
return self._module_registry.create_module(manifest=manifest)
instance()
classmethod
¶The default kiara context. In most cases, it's recommended you create and manage your own, though.
Source code in kiara/context/__init__.py
@classmethod
def instance(cls) -> "Kiara":
"""The default *kiara* context. In most cases, it's recommended you create and manage your own, though."""
if cls._instance is None:
cls._instance = Kiara()
return cls._instance
process(self, manifest, inputs)
¶Queue a job with the specified manifest and inputs.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
manifest |
Manifest |
the job manifest |
required |
inputs |
Mapping[str, Any] |
the job inputs |
required |
wait |
whether to wait for the job to be finished before returning |
required |
Returns
Source code in kiara/context/__init__.py
def process(self, manifest: Manifest, inputs: Mapping[str, Any]) -> ValueMap:
"""Queue a job with the specified manifest and inputs.
Arguments:
manifest: the job manifest
inputs: the job inputs
wait: whether to wait for the job to be finished before returning
Returns
"""
return self.job_registry.execute_and_retrieve(manifest=manifest, inputs=inputs)
queue(self, manifest, inputs, wait=False)
¶Queue a job with the specified manifest and inputs.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
manifest |
Manifest |
the job manifest |
required |
inputs |
Mapping[str, Any] |
the job inputs |
required |
wait |
bool |
whether to wait for the job to be finished before returning |
False |
Returns:
| Type | Description |
|---|---|
UUID |
the job id that can be used to look up job status & results |
Source code in kiara/context/__init__.py
def queue(
self, manifest: Manifest, inputs: Mapping[str, Any], wait: bool = False
) -> uuid.UUID:
"""Queue a job with the specified manifest and inputs.
Arguments:
manifest: the job manifest
inputs: the job inputs
wait: whether to wait for the job to be finished before returning
Returns:
the job id that can be used to look up job status & results
"""
return self.job_registry.execute(manifest=manifest, inputs=inputs, wait=wait)
save_values(self, values, alias_map)
¶Source code in kiara/context/__init__.py
def save_values(
self, values: ValueMap, alias_map: Mapping[str, Iterable[str]]
) -> StoreValuesResult:
_values = {}
for field_name in values.field_names:
value = values.get_value_obj(field_name)
_values[field_name] = value
self.data_registry.store_value(value=value, skip_if_exists=True)
stored = {}
for field_name, field_aliases in alias_map.items():
value = _values[field_name]
try:
if field_aliases:
self.alias_registry.register_aliases(value.value_id, *field_aliases)
stored[field_name] = StoreValueResult.construct(
value=value, aliases=sorted(field_aliases), error=None
)
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
stored[field_name] = StoreValueResult.construct(
value=value, aliases=sorted(field_aliases), error=str(e)
)
return StoreValuesResult.construct(__root__=stored)
KiaraContextInfo (KiaraModel)
pydantic-model
¶
Source code in kiara/context/__init__.py
class KiaraContextInfo(KiaraModel):
@classmethod
def create_from_kiara_instance(
cls, kiara: "Kiara", package_filter: Optional[str] = None
):
data_types = kiara.type_registry.get_context_metadata(
only_for_package=package_filter
)
modules = kiara.module_registry.get_context_metadata(
only_for_package=package_filter
)
operation_types = kiara.operation_registry.get_context_metadata(
only_for_package=package_filter
)
operations = filter_operations(
kiara=kiara, pkg_name=package_filter, **kiara.operation_registry.operations
)
metadata_types = find_metadata_models(only_for_package=package_filter)
return KiaraContextInfo.construct(
kiara_id=kiara.id,
package_filter=package_filter,
data_types=data_types,
module_types=modules,
metadata_types=metadata_types,
operation_types=operation_types,
operations=operations,
)
kiara_id: uuid.UUID = Field(description="The id of the kiara context.")
package_filter: Optional[str] = Field(
description="Whether this context is filtered to only include information included in a specific Python package."
)
data_types: DataTypeClassesInfo = Field(description="The included data types.")
module_types: ModuleTypeClassesInfo = Field(
description="The included kiara module types."
)
metadata_types: MetadataTypeClassesInfo = Field(
description="The included value metadata types."
)
operation_types: OperationTypeClassesInfo = Field(
description="The included operation types."
)
operations: OperationGroupInfo = Field(description="The included operations.")
def _retrieve_id(self) -> str:
if not self.package_filter:
return str(self.kiara_id)
else:
return f"{self.kiara_id}.package_{self.package_filter}"
def _retrieve_category_id(self) -> str:
return CONTEXT_INFO_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return {"kiara_id": self.kiara_id, "package": self.package_filter}
def get_info(self, item_type: str, item_id: str) -> KiaraInfoModel:
if "data_type" == item_type or "data_types" == item_type:
group_info: InfoModelGroup = self.data_types
elif "module" in item_type:
group_info = self.module_types
elif "metadata" in item_type:
group_info = self.metadata_types
elif "operation_type" in item_type or "operation_types" in item_type:
group_info = self.operation_types
elif "operation" in item_type:
group_info = self.operations
else:
item_types = [
"data_type",
"module_type",
"metadata_type",
"operation_type",
"operation",
]
raise Exception(
f"Can't determine item type '{item_type}', use one of: {', '.join(item_types)}"
)
return group_info[item_id]
def get_all_info(self, skip_empty_types: bool = True) -> Dict[str, InfoModelGroup]:
result: Dict[str, InfoModelGroup] = {}
if self.data_types or not skip_empty_types:
result["data_types"] = self.data_types
if self.module_types or not skip_empty_types:
result["module_types"] = self.module_types
if self.metadata_types or not skip_empty_types:
result["metadata_types"] = self.metadata_types
if self.operation_types or not skip_empty_types:
result["operation_types"] = self.operation_types
if self.operations or not skip_empty_types:
result["operations"] = self.operations
return result
Attributes¶
data_types: DataTypeClassesInfo
pydantic-field
required
¶The included data types.
kiara_id: UUID
pydantic-field
required
¶The id of the kiara context.
metadata_types: MetadataTypeClassesInfo
pydantic-field
required
¶The included value metadata types.
module_types: ModuleTypeClassesInfo
pydantic-field
required
¶The included kiara module types.
operation_types: OperationTypeClassesInfo
pydantic-field
required
¶The included operation types.
operations: OperationGroupInfo
pydantic-field
required
¶The included operations.
package_filter: str
pydantic-field
¶Whether this context is filtered to only include information included in a specific Python package.
create_from_kiara_instance(kiara, package_filter=None)
classmethod
¶Source code in kiara/context/__init__.py
@classmethod
def create_from_kiara_instance(
cls, kiara: "Kiara", package_filter: Optional[str] = None
):
data_types = kiara.type_registry.get_context_metadata(
only_for_package=package_filter
)
modules = kiara.module_registry.get_context_metadata(
only_for_package=package_filter
)
operation_types = kiara.operation_registry.get_context_metadata(
only_for_package=package_filter
)
operations = filter_operations(
kiara=kiara, pkg_name=package_filter, **kiara.operation_registry.operations
)
metadata_types = find_metadata_models(only_for_package=package_filter)
return KiaraContextInfo.construct(
kiara_id=kiara.id,
package_filter=package_filter,
data_types=data_types,
module_types=modules,
metadata_types=metadata_types,
operation_types=operation_types,
operations=operations,
)
get_all_info(self, skip_empty_types=True)
¶Source code in kiara/context/__init__.py
def get_all_info(self, skip_empty_types: bool = True) -> Dict[str, InfoModelGroup]:
result: Dict[str, InfoModelGroup] = {}
if self.data_types or not skip_empty_types:
result["data_types"] = self.data_types
if self.module_types or not skip_empty_types:
result["module_types"] = self.module_types
if self.metadata_types or not skip_empty_types:
result["metadata_types"] = self.metadata_types
if self.operation_types or not skip_empty_types:
result["operation_types"] = self.operation_types
if self.operations or not skip_empty_types:
result["operations"] = self.operations
return result
get_info(self, item_type, item_id)
¶Source code in kiara/context/__init__.py
def get_info(self, item_type: str, item_id: str) -> KiaraInfoModel:
if "data_type" == item_type or "data_types" == item_type:
group_info: InfoModelGroup = self.data_types
elif "module" in item_type:
group_info = self.module_types
elif "metadata" in item_type:
group_info = self.metadata_types
elif "operation_type" in item_type or "operation_types" in item_type:
group_info = self.operation_types
elif "operation" in item_type:
group_info = self.operations
else:
item_types = [
"data_type",
"module_type",
"metadata_type",
"operation_type",
"operation",
]
raise Exception(
f"Can't determine item type '{item_type}', use one of: {', '.join(item_types)}"
)
return group_info[item_id]
Functions¶
explain(item)
¶
Pretty print information about an item on the terminal.
Source code in kiara/context/__init__.py
def explain(item: Any):
"""Pretty print information about an item on the terminal."""
if isinstance(item, type):
from kiara.modules import KiaraModule
if issubclass(item, KiaraModule):
item = KiaraModuleTypeInfo.create_from_type_class(type_cls=item)
console = get_console()
console.print(item)
Modules¶
config
¶
yaml
¶Classes¶
KiaraArchiveConfig (BaseModel)
pydantic-model
¶Source code in kiara/context/config.py
class KiaraArchiveConfig(BaseModel):
archive_id: str = Field(description="The unique archive id.")
archive_type: str = Field(description="The archive type.")
config: Mapping[str, Any] = Field(
description="Archive type specific config.", default_factory=dict
)
@property
def archive_uuid(self) -> uuid.UUID:
return uuid.UUID(self.archive_id)
KiaraBaseConfig (BaseSettings)
pydantic-model
¶Source code in kiara/context/config.py
class KiaraBaseConfig(BaseSettings):
class Config:
extra = Extra.forbid
archives: Dict[str, KiaraArchiveConfig] = Field(
description="All the archives this kiara context can use and the aliases they are registered with."
)
extra_pipeline_folders: List[str] = Field(
description="Paths to local folders that contain kiara pipelines.",
default_factory=list,
)
ignore_errors: bool = Field(
description="If set, kiara will try to ignore most errors (that can be ignored).",
default=False,
)
archives: Dict[str, kiara.context.config.KiaraArchiveConfig]
pydantic-field
required
¶All the archives this kiara context can use and the aliases they are registered with.
extra_pipeline_folders: List[str]
pydantic-field
¶Paths to local folders that contain kiara pipelines.
ignore_errors: bool
pydantic-field
¶If set, kiara will try to ignore most errors (that can be ignored).
Config
¶Source code in kiara/context/config.py
class Config:
extra = Extra.forbid
KiaraConfig (BaseSettings)
pydantic-model
¶Source code in kiara/context/config.py
class KiaraConfig(BaseSettings):
class Config:
env_prefix = "kiara_"
extra = Extra.forbid
@classmethod
def customise_sources(
cls,
init_settings,
env_settings,
file_secret_settings,
):
return (init_settings, env_settings, config_file_settings_source)
base_path: str = Field(
description="The base path to use for all data (unless otherwise specified.",
default=kiara_app_dirs.user_data_dir,
)
context_base_path: str = Field(description="The base path to look for contexts in.")
stores_base_path: str = Field(
description="The base path for the stores of this context."
)
default_context: str = Field(
description="The name of the default context to use if none is provided.",
default=DEFAULT_CONTEXT_NAME,
)
@root_validator(pre=True)
def _set_paths(cls, values: Any):
base_path = values.get("base_path", None)
if not base_path:
base_path = kiara_app_dirs.user_data_dir
values["base_path"] = base_path
elif isinstance(base_path, Path):
base_path = base_path.as_posix()
context_base_path = values.get("context_base_path", None)
if not context_base_path:
context_base_path = os.path.join(base_path, "contexts")
values["context_base_path"] = context_base_path
stores_base_path = values.get("stores_base_path", None)
if not stores_base_path:
stores_base_path = os.path.join(base_path, "stores")
values["stores_base_path"] = stores_base_path
return values
def create_context_config(
self, context_name: Optional[str] = None, **extra_args: Any
) -> "KiaraCurrentContextConfig":
return KiaraCurrentContextConfig(
kiara_config=self, context=context_name, **extra_args
)
def create_context(
self, context_name: Optional[str] = None, **extra_args: Any
) -> "Kiara":
config = self.create_context_config(context_name=context_name, **extra_args)
from kiara.context import Kiara
return Kiara(config=config.get_context())
base_path: str
pydantic-field
¶The base path to use for all data (unless otherwise specified.
context_base_path: str
pydantic-field
required
¶The base path to look for contexts in.
default_context: str
pydantic-field
¶The name of the default context to use if none is provided.
stores_base_path: str
pydantic-field
required
¶The base path for the stores of this context.
Config
¶Source code in kiara/context/config.py
class Config:
env_prefix = "kiara_"
extra = Extra.forbid
@classmethod
def customise_sources(
cls,
init_settings,
env_settings,
file_secret_settings,
):
return (init_settings, env_settings, config_file_settings_source)
env_prefix
¶extra
¶customise_sources(init_settings, env_settings, file_secret_settings)
classmethod
¶Source code in kiara/context/config.py
@classmethod
def customise_sources(
cls,
init_settings,
env_settings,
file_secret_settings,
):
return (init_settings, env_settings, config_file_settings_source)
create_context(self, context_name=None, **extra_args)
¶Source code in kiara/context/config.py
def create_context(
self, context_name: Optional[str] = None, **extra_args: Any
) -> "Kiara":
config = self.create_context_config(context_name=context_name, **extra_args)
from kiara.context import Kiara
return Kiara(config=config.get_context())
create_context_config(self, context_name=None, **extra_args)
¶Source code in kiara/context/config.py
def create_context_config(
self, context_name: Optional[str] = None, **extra_args: Any
) -> "KiaraCurrentContextConfig":
return KiaraCurrentContextConfig(
kiara_config=self, context=context_name, **extra_args
)
KiaraContextConfig (KiaraBaseConfig)
pydantic-model
¶Source code in kiara/context/config.py
class KiaraContextConfig(KiaraBaseConfig):
class Config:
extra = Extra.forbid
context_id: str = Field(description="A globally unique id for this kiara context.")
context_alias: str = Field(
description="An alias for this kiara context, must be unique within all known contexts."
)
context_folder: str = Field(
description="The base folder where settings and data for this kiara context will be stored."
)
@property
def db_url(self):
return get_kiara_db_url(self.context_folder)
@property
def data_directory(self) -> str:
return os.path.join(self.context_folder, "data")
context_alias: str
pydantic-field
required
¶An alias for this kiara context, must be unique within all known contexts.
context_folder: str
pydantic-field
required
¶The base folder where settings and data for this kiara context will be stored.
context_id: str
pydantic-field
required
¶A globally unique id for this kiara context.
data_directory: str
property
readonly
¶db_url
property
readonly
¶
Config
¶Source code in kiara/context/config.py
class Config:
extra = Extra.forbid
KiaraCurrentContextConfig (KiaraBaseConfig)
pydantic-model
¶Configuration that holds the currently active context, as well as references to other available contexts.
Source code in kiara/context/config.py
class KiaraCurrentContextConfig(KiaraBaseConfig):
"""Configuration that holds the currently active context, as well as references to other available contexts."""
class Config:
env_prefix = "kiara_context_"
extra = Extra.forbid
@classmethod
def customise_sources(
cls,
init_settings,
env_settings,
file_secret_settings,
):
return (
init_settings,
env_settings,
)
kiara_config: KiaraConfig = Field(
description="The base kiara configuration.", default_factory=KiaraConfig
)
context: str = Field(
description=f"The path to an existing folder that houses the context, or the name of the context to use under the default kiara app data directory ({kiara_app_dirs.user_data_dir})."
)
context_configs: Dict[str, KiaraContextConfig] = Field(
description="The context configuration."
)
# overlay_config: KiaraConfig = Field(description="Extra config options to add to the selected context.")
@classmethod
def find_current_contexts(
cls, kiara_config: KiaraConfig
) -> Dict[str, KiaraContextConfig]:
contexts: Dict[str, KiaraContextConfig] = {}
if not os.path.exists(kiara_config.context_base_path):
return contexts
for f in os.listdir(kiara_config.context_base_path):
config_dir = os.path.join(kiara_config.context_base_path, f)
k_config = cls.load_context(config_dir)
if k_config:
contexts[k_config.context_alias] = k_config
return contexts
@classmethod
def create_context(
cls,
path: str,
context_id: str,
kiara_config: KiaraConfig,
context_alias: Optional[str],
) -> KiaraContextConfig:
if os.path.exists(path):
raise Exception(f"Can't create kiara context folder, path exists: {path}")
os.makedirs(path, exist_ok=False)
config = {}
config["context_id"] = context_id
if not context_alias:
context_alias = config["context_id"]
config["context_alias"] = context_alias
config["context_folder"] = path
config["archives"] = create_default_archives(kiara_config=kiara_config)
kiara_context_config = KiaraContextConfig(**config)
config_file = os.path.join(path, "kiara_context.yaml")
with open(config_file, "wt") as f:
yaml.dump(kiara_config.dict(), f)
return kiara_context_config
@classmethod
def load_context(cls, path: str):
if path.endswith("kiara_context.yaml"):
path = os.path.dirname(path)
if not os.path.isdir(path):
return None
config_file = os.path.join(path, "kiara_context.yaml")
if not os.path.isfile(config_file):
return None
try:
config = get_data_from_file(config_file)
k_config = KiaraContextConfig(**config)
except Exception as e:
log_message("config.parse.error", config_file=config_file, error=e)
return None
return k_config
@root_validator(pre=True)
def validate_global_config(cls, values):
create_context = values.pop("create_context", False)
kiara_config = values.get("kiara_config", None)
if kiara_config is None:
kiara_config = KiaraConfig()
values["kiara_config"] = kiara_config
contexts = cls.find_current_contexts(kiara_config=kiara_config)
assert "context_configs" not in values.keys()
assert "overlay_config" not in values.keys()
context_name: Optional[str] = values.get("context", None)
if context_name is None:
context_name = kiara_config.default_context
loaded_context: Optional[KiaraContextConfig] = None
assert context_name != "kiara_context.yaml"
if context_name != DEFAULT_CONTEXT_NAME:
context_dir: Optional[str] = None
if context_name.endswith("kiara_context.yaml"):
context_dir = os.path.dirname(context_name)
elif os.path.isdir(context_name):
context_config = os.path.join(context_name, "kiara_context.yaml")
if os.path.exists(context_config):
context_dir = context_name
if context_dir is not None:
loaded_context = loaded_context(context_dir)
elif create_context and os.path.sep in context_name:
# we assume this is meant to be a path that is outside of the 'normal' kiara data directory
if context_name.endswith("kiara_context.yaml"):
context_dir = os.path.dirname(context_name)
else:
context_dir = os.path.abspath(os.path.expanduser(context_name))
context_id = str(uuid.uuid4())
loaded_context = cls.create_context(
path=context_dir, context_id=context_id, kiara_config=kiara_config
)
if loaded_context is not None:
contexts[loaded_context.context_alias] = loaded_context
context_name = loaded_context.context_alias
else:
match = None
for context_alias, context in contexts.items():
if context.context_id == context_name:
if match:
raise Exception(
f"More then one kiara contexts with id: {context.context_id}"
)
match = context_name
elif context.context_alias == context_name:
if match:
raise Exception(
f"More then one kiara contexts with alias: {context.context_id}"
)
match = context_name
if not match:
if not create_context and context_name != DEFAULT_CONTEXT_NAME:
raise Exception(f"Can't find context with name: {context_name}")
context_id = str(uuid.uuid4())
context_dir = os.path.join(kiara_config.context_base_path, context_id)
kiara_config = cls.create_context(
path=context_dir,
context_id=context_id,
context_alias=context_name,
kiara_config=kiara_config,
)
contexts[context_name] = kiara_config
else:
context_name = match
values["context"] = context_name
values["context_configs"] = contexts
values["archives"] = contexts[context_name].archives
return values
def get_context(self, context_name: Optional[str] = None) -> KiaraContextConfig:
if not context_name:
context_name = self.context
if context_name not in self.context_configs.keys():
raise Exception(
f"Kiara context '{context_name}' not registered. Registered contexts: {', '.join(self.context_configs.keys())}"
)
selected_dict = self.context_configs[context_name].dict()
overlay = self.dict(exclude={"context", "context_configs", "kiara_config"})
selected_dict.update(overlay)
kc = KiaraContextConfig(**selected_dict)
return kc
def create_renderable(self, **config) -> RenderableType:
return create_table_from_model_object(self)
context: str
pydantic-field
required
¶The path to an existing folder that houses the context, or the name of the context to use under the default kiara app data directory (/home/runner/.local/share/kiara).
context_configs: Dict[str, kiara.context.config.KiaraContextConfig]
pydantic-field
required
¶The context configuration.
kiara_config: KiaraConfig
pydantic-field
¶The base kiara configuration.
Config
¶Source code in kiara/context/config.py
class Config:
env_prefix = "kiara_context_"
extra = Extra.forbid
@classmethod
def customise_sources(
cls,
init_settings,
env_settings,
file_secret_settings,
):
return (
init_settings,
env_settings,
)
create_context(path, context_id, kiara_config, context_alias)
classmethod
¶Source code in kiara/context/config.py
@classmethod
def create_context(
cls,
path: str,
context_id: str,
kiara_config: KiaraConfig,
context_alias: Optional[str],
) -> KiaraContextConfig:
if os.path.exists(path):
raise Exception(f"Can't create kiara context folder, path exists: {path}")
os.makedirs(path, exist_ok=False)
config = {}
config["context_id"] = context_id
if not context_alias:
context_alias = config["context_id"]
config["context_alias"] = context_alias
config["context_folder"] = path
config["archives"] = create_default_archives(kiara_config=kiara_config)
kiara_context_config = KiaraContextConfig(**config)
config_file = os.path.join(path, "kiara_context.yaml")
with open(config_file, "wt") as f:
yaml.dump(kiara_config.dict(), f)
return kiara_context_config
create_renderable(self, **config)
¶Source code in kiara/context/config.py
def create_renderable(self, **config) -> RenderableType:
return create_table_from_model_object(self)
find_current_contexts(kiara_config)
classmethod
¶Source code in kiara/context/config.py
@classmethod
def find_current_contexts(
cls, kiara_config: KiaraConfig
) -> Dict[str, KiaraContextConfig]:
contexts: Dict[str, KiaraContextConfig] = {}
if not os.path.exists(kiara_config.context_base_path):
return contexts
for f in os.listdir(kiara_config.context_base_path):
config_dir = os.path.join(kiara_config.context_base_path, f)
k_config = cls.load_context(config_dir)
if k_config:
contexts[k_config.context_alias] = k_config
return contexts
get_context(self, context_name=None)
¶Source code in kiara/context/config.py
def get_context(self, context_name: Optional[str] = None) -> KiaraContextConfig:
if not context_name:
context_name = self.context
if context_name not in self.context_configs.keys():
raise Exception(
f"Kiara context '{context_name}' not registered. Registered contexts: {', '.join(self.context_configs.keys())}"
)
selected_dict = self.context_configs[context_name].dict()
overlay = self.dict(exclude={"context", "context_configs", "kiara_config"})
selected_dict.update(overlay)
kc = KiaraContextConfig(**selected_dict)
return kc
load_context(path)
classmethod
¶Source code in kiara/context/config.py
@classmethod
def load_context(cls, path: str):
if path.endswith("kiara_context.yaml"):
path = os.path.dirname(path)
if not os.path.isdir(path):
return None
config_file = os.path.join(path, "kiara_context.yaml")
if not os.path.isfile(config_file):
return None
try:
config = get_data_from_file(config_file)
k_config = KiaraContextConfig(**config)
except Exception as e:
log_message("config.parse.error", config_file=config_file, error=e)
return None
return k_config
validate_global_config(values)
classmethod
¶Source code in kiara/context/config.py
@root_validator(pre=True)
def validate_global_config(cls, values):
create_context = values.pop("create_context", False)
kiara_config = values.get("kiara_config", None)
if kiara_config is None:
kiara_config = KiaraConfig()
values["kiara_config"] = kiara_config
contexts = cls.find_current_contexts(kiara_config=kiara_config)
assert "context_configs" not in values.keys()
assert "overlay_config" not in values.keys()
context_name: Optional[str] = values.get("context", None)
if context_name is None:
context_name = kiara_config.default_context
loaded_context: Optional[KiaraContextConfig] = None
assert context_name != "kiara_context.yaml"
if context_name != DEFAULT_CONTEXT_NAME:
context_dir: Optional[str] = None
if context_name.endswith("kiara_context.yaml"):
context_dir = os.path.dirname(context_name)
elif os.path.isdir(context_name):
context_config = os.path.join(context_name, "kiara_context.yaml")
if os.path.exists(context_config):
context_dir = context_name
if context_dir is not None:
loaded_context = loaded_context(context_dir)
elif create_context and os.path.sep in context_name:
# we assume this is meant to be a path that is outside of the 'normal' kiara data directory
if context_name.endswith("kiara_context.yaml"):
context_dir = os.path.dirname(context_name)
else:
context_dir = os.path.abspath(os.path.expanduser(context_name))
context_id = str(uuid.uuid4())
loaded_context = cls.create_context(
path=context_dir, context_id=context_id, kiara_config=kiara_config
)
if loaded_context is not None:
contexts[loaded_context.context_alias] = loaded_context
context_name = loaded_context.context_alias
else:
match = None
for context_alias, context in contexts.items():
if context.context_id == context_name:
if match:
raise Exception(
f"More then one kiara contexts with id: {context.context_id}"
)
match = context_name
elif context.context_alias == context_name:
if match:
raise Exception(
f"More then one kiara contexts with alias: {context.context_id}"
)
match = context_name
if not match:
if not create_context and context_name != DEFAULT_CONTEXT_NAME:
raise Exception(f"Can't find context with name: {context_name}")
context_id = str(uuid.uuid4())
context_dir = os.path.join(kiara_config.context_base_path, context_id)
kiara_config = cls.create_context(
path=context_dir,
context_id=context_id,
context_alias=context_name,
kiara_config=kiara_config,
)
contexts[context_name] = kiara_config
else:
context_name = match
values["context"] = context_name
values["context_configs"] = contexts
values["archives"] = contexts[context_name].archives
return values
config_file_settings_source(settings)
¶Source code in kiara/context/config.py
def config_file_settings_source(settings: BaseSettings) -> Dict[str, Any]:
if os.path.isfile(KIARA_MAIN_CONFIG_FILE):
config = get_data_from_file(KIARA_MAIN_CONFIG_FILE)
if not isinstance(config, Mapping):
raise ValueError(
f"Invalid config file format, can't parse file: {KIARA_MAIN_CONFIG_FILE}"
)
else:
config = {}
return config
create_default_archives(kiara_config)
¶Source code in kiara/context/config.py
def create_default_archives(kiara_config: KiaraConfig):
env_registry = EnvironmentRegistry.instance()
archives = env_registry.environments["kiara_types"].archive_types
data_store_type = "filesystem_data_store"
assert data_store_type in archives.keys()
data_store_id = ID_REGISTRY.generate(comment="default data store id")
data_archive_config = {
"base_path": os.path.join(kiara_config.stores_base_path, data_store_type)
}
data_store = KiaraArchiveConfig.construct(
archive_id=str(data_store_id),
archive_type=data_store_type,
config=data_archive_config,
)
job_store_type = "filesystem_job_store"
job_archive_config = {
"base_path": os.path.join(kiara_config.stores_base_path, job_store_type)
}
job_store_id = ID_REGISTRY.generate(comment="default job store id")
job_store = KiaraArchiveConfig.construct(
archive_id=str(job_store_id),
archive_type=job_store_type,
config=job_archive_config,
)
alias_store_type = "filesystem_alias_store"
alias_store_config = {
"base_path": os.path.join(kiara_config.stores_base_path, alias_store_type)
}
alias_store_id = ID_REGISTRY.generate(comment="default alias store id")
alias_store = KiaraArchiveConfig.construct(
archive_id=str(alias_store_id),
archive_type=alias_store_type,
config=alias_store_config,
)
return {
DEFAULT_DATA_STORE_MARKER: data_store,
DEFAULT_JOB_STORE_MARKER: job_store,
DEFAULT_ALIAS_STORE_MARKER: alias_store,
}
orm
¶
Base: DeclarativeMeta
¶jobs_env_association_table
¶value_env_association_table
¶
AliasOrm (Base)
¶Source code in kiara/context/orm.py
class AliasOrm(Base):
__tablename__ = "aliases"
id: Column[Optional[int]] = Column(Integer, primary_key=True)
alias: Column[str] = Column(String, index=True, nullable=False)
created: Column[datetime] = Column(UtcDateTime(), nullable=False, index=True)
version: Column[int] = Column(Integer, nullable=False, index=True)
value_id: Column[Optional[uuid.UUID]] = Column(UUIDType(binary=True), nullable=True)
UniqueConstraint(alias, version)
DestinyOrm (Base)
¶Source code in kiara/context/orm.py
class DestinyOrm(Base):
__tablename__ = "destinies"
id: Column[Optional[int]] = Column(Integer, primary_key=True)
value_id: Column[int] = Column(Integer, ForeignKey("values.id"), nullable=False)
category: Column[str] = Column(String, nullable=False, index=False)
key: Column[str] = Column(String, nullable=False, index=False)
manifest_id: Column[int] = Column(
Integer, ForeignKey("manifests.id"), nullable=False
)
inputs: Column[Union[Dict[Any, Any], List[Any]]] = Column(
JSON, index=False, nullable=False
)
output_name: Column[str] = Column(String, index=False, nullable=False)
destiny_value: Column[Optional[int]] = Column(
Integer, ForeignKey("values.id"), nullable=True
)
description: Column[Optional[str]] = Column(String, nullable=True)
UniqueConstraint(value_id, category, key)
EnvironmentOrm (Base)
¶Source code in kiara/context/orm.py
class EnvironmentOrm(Base):
__tablename__ = "environments"
id: Column[Optional[int]] = Column(Integer, primary_key=True)
metadata_hash: Column[int] = Column(Integer, index=True, nullable=False)
metadata_schema_id = Column(
Integer, ForeignKey("metadata_schema_lookup.id"), nullable=False
)
metadata_payload: Column[Union[Dict[Any, Any], List[Any]]] = Column(
JSON, nullable=False
)
UniqueConstraint(metadata_hash)
JobsOrm (Base)
¶Source code in kiara/context/orm.py
class JobsOrm(Base):
__tablename__ = "jobs"
id: Column[Optional[int]] = Column(Integer, primary_key=True)
manifest_id: Column[int] = Column(
Integer, ForeignKey("manifests.id"), nullable=False
)
inputs: Column[Union[Dict[Any, Any], List[Any]]] = Column(JSON, nullable=False)
input_hash: Column[str] = Column(String, nullable=False)
is_idempotent: Column[bool] = Column(Boolean, nullable=False)
created: Column[datetime] = Column(UtcDateTime(), default=utcnow(), nullable=False)
started: Column[Optional[datetime]] = Column(UtcDateTime(), nullable=True)
duration_ms: Column[Optional[int]] = Column(Integer, nullable=True)
environments = relationship("EnvironmentOrm", secondary=jobs_env_association_table)
ManifestOrm (Base)
¶Source code in kiara/context/orm.py
class ManifestOrm(Base):
__tablename__ = "manifests"
id: Column[Optional[int]] = Column(Integer, primary_key=True)
module_type: Column[str] = Column(String, index=True, nullable=False)
module_config: Column[Union[Dict[Any, Any], List[Any]]] = Column(
JSON, nullable=False
)
manifest_hash: Column[int] = Column(Integer, index=True, nullable=False)
is_idempotent: Column[bool] = Column(Boolean, nullable=False)
UniqueConstraint(module_type, manifest_hash)
MetadataSchemaOrm (Base)
¶Source code in kiara/context/orm.py
class MetadataSchemaOrm(Base):
__tablename__ = "metadata_schema_lookup"
id: Column[Optional[int]] = Column(Integer, primary_key=True)
metadata_schema_hash: Column[int] = Column(Integer, index=True, nullable=False)
metadata_type: Column[str] = Column(String, nullable=False)
metadata_schema: Column[Union[Dict[Any, Any], List[Any]]] = Column(
JSON, nullable=False
)
metadata_payloads = relationship("EnvironmentOrm")
UniqueConstraint(metadata_schema_hash)
Pedigree (Base)
¶Source code in kiara/context/orm.py
class Pedigree(Base):
__tablename__ = "pedigrees"
id: Column[Optional[int]] = Column(Integer, primary_key=True)
manifest_id: Column[int] = Column(
Integer, ForeignKey("manifests.id"), nullable=False
)
inputs: Column[Union[Dict[Any, Any], List[Any]]] = Column(JSON, nullable=False)
ValueOrm (Base)
¶Source code in kiara/context/orm.py
class ValueOrm(Base):
__tablename__ = "values"
id: Column[Optional[int]] = Column(Integer, primary_key=True)
global_id: Column[uuid.UUID] = Column(UUIDType(binary=True), nullable=False)
data_type_id: Column[int] = Column(
Integer, ForeignKey("data_types.id"), nullable=False
)
data_type_name: Column[str] = Column(String, index=True, nullable=False)
value_size: Column[int] = Column(Integer, index=True, nullable=False)
value_hash: Column[str] = Column(String, index=True, nullable=False)
environments = relationship("EnvironmentOrm", secondary=value_env_association_table)
UniqueConstraint(value_hash, value_size, data_type_id)
ValueTypeOrm (Base)
¶Source code in kiara/context/orm.py
class ValueTypeOrm(Base):
__tablename__ = "data_types"
id: Column[Optional[int]] = Column(Integer, primary_key=True)
type_config_hash: Column[int] = Column(Integer, index=True, nullable=False)
type_name: Column[str] = Column(String, nullable=False, index=True)
type_config: Column[Union[Dict[Any, Any], List[Any]]] = Column(JSON, nullable=False)
UniqueConstraint(type_config_hash, type_name)
data_types
special
¶
This is the base module that contains everything data type-related in kiara.
I'm still not 100% sure how to best implement the kiara type system, there are several ways it could be done, for example based on Python type-hints, using JSON-schema, Avro (which is my 2nd favourite option), as well as by implementing a custom type-class hierarchy. Which is what I have choosen to try first. For now, it looks like it'll work out, but there is a chance requirements I haven't forseen will crop up that could make this become ugly.
Anyway, the way it works (for now) is that kiara comes with a set of often used data_types (the standard set of: scalars, list, dict, table & array, etc.) which each come with 2 functions that can serialize and deserialize values of that type in a persistant fashion -- which could be storing as a file on disk, or as a cell/row in a database. Those functions will most likley be kiara modules themselves, with even more restricted input/output type options.
In addition, packages that contain modules can implement their own, custom data_types, if suitable ones are not available in core-kiara. Those can either be 'serialized/deserialized' into kiara-native data_types (which in turn will serialize them using their own serializing functions), or will have to implement custom serializing functionality (which will probably be discouraged, since this might not be trivial and there are quite a few things to consider).
TYPE_CONFIG_CLS
¶
TYPE_PYTHON_CLS
¶
Classes¶
DataType (ABC, Generic)
¶
Base class that all kiara data_types must inherit from.
kiara data_types have 3 main responsibilities:
- serialize into / deserialize from persistent state
- data validation
- metadata extraction
Serializing being the arguably most important of those, because without most of the data management features of kiara would be impossible. Validation should not require any explanation. Metadata extraction is important, because that metadata will be available to other components of kiara (or frontends for it), without them having to request the actual data. That will hopefully make kiara very efficient in terms of memory management, as well as data transfer and I/O. Ideally, the actual data (bytes) will only be requested at the last possible moment. For example when a module needs the input data to do processing on it -- and even then it might be that it only requests a part of the data, say a single column of a table. Or when a frontend needs to display/visualize the data.
Source code in kiara/data_types/__init__.py
class DataType(abc.ABC, Generic[TYPE_PYTHON_CLS, TYPE_CONFIG_CLS]):
"""Base class that all *kiara* data_types must inherit from.
*kiara* data_types have 3 main responsibilities:
- serialize into / deserialize from persistent state
- data validation
- metadata extraction
Serializing being the arguably most important of those, because without most of the data management features of
*kiara* would be impossible. Validation should not require any explanation. Metadata extraction is important, because
that metadata will be available to other components of *kiara* (or frontends for it), without them having to request
the actual data. That will hopefully make *kiara* very efficient in terms of memory management, as well as data
transfer and I/O. Ideally, the actual data (bytes) will only be requested at the last possible moment. For example when a
module needs the input data to do processing on it -- and even then it might be that it only requests a part of the
data, say a single column of a table. Or when a frontend needs to display/visualize the data.
"""
@classmethod
def retrieve_available_type_profiles(cls) -> Mapping[str, Mapping[str, Any]]:
return {}
@classmethod
@abc.abstractmethod
def python_class(cls) -> Type[TYPE_PYTHON_CLS]:
pass
@classmethod
def data_type_config_class(cls) -> Type[TYPE_CONFIG_CLS]:
return DataTypeConfig # type: ignore
@classmethod
def _calculate_data_type_hash(
cls, data_type_config: Union[Mapping[str, Any], DataTypeConfig]
) -> int:
if isinstance(data_type_config, Mapping):
data_type_config = cls.data_type_config_class()(**data_type_config) # type: ignore
obj = {
"type": cls._data_type_name, # type: ignore
"type_config": data_type_config.config_hash,
}
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
return h[obj]
def __init__(self, **type_config: Any):
try:
self._type_config: TYPE_CONFIG_CLS = self.__class__.data_type_config_class()(**type_config) # type: ignore # TODO: double-check this is only a mypy issue
except ValidationError as ve:
raise ValueTypeConfigException(
f"Error creating object for type: {ve}",
self.__class__,
type_config,
ve,
)
self._data_type_hash: Optional[int] = None
self._characteristics: Optional[DataTypeCharacteristics] = None
@property
def data_type_name(self) -> str:
return self._data_type_name # type: ignore
@property
def data_type_hash(self) -> int:
if self._data_type_hash is None:
self._data_type_hash = self.__class__._calculate_data_type_hash(
self._type_config
)
return self._data_type_hash
@property
def characteristics(self) -> DataTypeCharacteristics:
if self._characteristics is not None:
return self._characteristics
self._characteristics = self._retrieve_characteristics()
return self._characteristics
def _retrieve_characteristics(self) -> DataTypeCharacteristics:
return DataTypeCharacteristics()
# @abc.abstractmethod
# def is_immutable(self) -> bool:
# pass
@abc.abstractmethod
def calculate_hash(self, data: TYPE_PYTHON_CLS) -> int:
"""Calculate the hash of the value."""
@abc.abstractmethod
def calculate_size(self, data: TYPE_PYTHON_CLS) -> int:
pass
@property
def type_config(self) -> TYPE_CONFIG_CLS:
return self._type_config
def _pre_examine_data(
self, data: Any, schema: ValueSchema
) -> Tuple[Any, ValueStatus, int]:
if data == SpecialValue.NOT_SET:
status = ValueStatus.NOT_SET
else:
status = ValueStatus.SET
if data is None:
if schema.default in [None, SpecialValue.NO_VALUE]:
data = SpecialValue.NO_VALUE
status = ValueStatus.NONE
elif schema.default == SpecialValue.NOT_SET:
data = SpecialValue.NOT_SET
status = ValueStatus.NOT_SET
elif callable(schema.default):
data = schema.default()
status = ValueStatus.DEFAULT
else:
data = copy.deepcopy(schema.default)
status = ValueStatus.DEFAULT
else:
data = self.parse_python_obj(data)
if data is None:
raise Exception(
f"Invalid data, can't parse into a value of type '{schema.type}'."
)
if status in [ValueStatus.SET, ValueStatus.DEFAULT]:
value_hash = self.calculate_hash(data)
else:
value_hash = 0
return (data, status, value_hash)
# def create_value(self, data: Any, schema: Optional[ValueSchema]=None, pedigree: Optional[ValuePedigree]=None) -> TYPE_VALUE_CLS:
#
# if schema is None:
# raise NotImplementedError()
#
# if pedigree is None:
# raise NotImplementedError()
#
# data, status, value_hash = self._pre_examine_data(data=data, schema=schema)
#
# v_id = uuid.uuid4()
# value = self.assemble_value(value_id=v_id, data=data, schema=schema, status=status, value_hash=value_hash, pedigree=pedigree, kiara_id=kiara_id)
# return value
# def reassemble_value(self, value_id: uuid.UUID, load_config: "LoadConfig", schema: ValueSchema, status: Union[ValueStatus, str],
# value_hash: int, value_size: int, pedigree: ValuePedigree, kiara_id: uuid.UUID,
# pedigree_output_name: str) -> TYPE_VALUE_CLS:
#
# if isinstance(status, str):
# status = ValueStatus(status)
#
# value_cls = self.value_class()
# if status in [ValueStatus.SET, ValueStatus.DEFAULT]:
#
# try:
#
# value = value_cls(
# value_id=value_id,
# kiara_id=kiara_id,
# value_status=status,
# value_size=value_size,
# value_hash=value_hash,
# value_schema=schema,
# pedigree=pedigree,
# pedigree_output_name=pedigree_output_name
# )
#
# except Exception as e:
# raise KiaraValueException(
# data_type=self.__class__, value_data=load_config, exception=e
# )
# retriever = LoadConfigRetriever(load_config=load_config)
# value._data_retriever = retriever
#
# else:
# assert value_size == 0
# value = value_cls(
# value_id=value_id,
# kiara_id=kiara_id,
# value_status=status,
# value_size=value_size,
# value_hash=value_hash,
# value_schema=schema,
# pedigree=pedigree,
# pedigree_output_name=pedigree_output_name
# )
#
# if status == ValueStatus.NONE:
# data = SpecialValue.NO_VALUE
# else:
# data = SpecialValue.NOT_SET
#
# retriever = StaticDataRetriever(data=data)
# value._data_retriever = retriever
#
# return value
def assemble_value(
self,
value_id: uuid.UUID,
data: Any,
schema: ValueSchema,
status: Union[ValueStatus, str],
value_hash: int,
pedigree: ValuePedigree,
kiara_id: uuid.UUID,
pedigree_output_name: str,
) -> Tuple[Value, Any]:
if isinstance(status, str):
status = ValueStatus(status).name
this_cls = PythonClass.from_class(self.__class__)
if status in [ValueStatus.SET, ValueStatus.DEFAULT]:
size = self.calculate_size(data)
try:
self._validate(data)
value = Value(
value_id=value_id,
kiara_id=kiara_id,
value_status=status,
value_size=size,
value_hash=value_hash,
value_schema=schema,
pedigree=pedigree,
pedigree_output_name=pedigree_output_name,
data_type_class=this_cls,
)
except Exception as e:
raise KiaraValueException(
data_type=self.__class__, value_data=data, exception=e
)
else:
size = 0
value = Value(
value_id=value_id,
kiara_id=kiara_id,
value_status=status,
value_size=size,
value_hash=value_hash,
value_schema=schema,
pedigree=pedigree,
pedigree_output_name=pedigree_output_name,
data_type_class=this_cls,
)
value._value_data = data
value._data_type = self
return value, data
def parse_python_obj(self, data: Any) -> TYPE_PYTHON_CLS:
"""Parse a value into a supported python type.
This exists to make it easier to do trivial conversions (e.g. from a date string to a datetime object).
If you choose to overwrite this method, make 100% sure that you don't change the meaning of the value, and try to
avoid adding or removing information from the data (e.g. by changing the resolution of a date).
Arguments:
v: the value
Returns:
'None', if no parsing was done and the original value should be used, otherwise return the parsed Python object
"""
return data
def _validate(self, value: TYPE_PYTHON_CLS) -> None:
"""Validate the value. This expects an instance of the defined Python class (from 'backing_python_type)."""
if not isinstance(value, self.__class__.python_class()):
raise ValueError(
f"Invalid python type '{type(value)}', must be: {self.__class__.python_class()}"
)
def create_renderable(self, **config):
show_type_info = config.get("show_type_info", False)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("key")
table.add_column("value", style="i")
table.add_row("type_name", self.data_type_name)
config_json = self.type_config.json(
exclude_unset=True, option=orjson.OPT_INDENT_2
)
config = Syntax(config_json, "json", background_color="default")
table.add_row("type_config", config)
if show_type_info:
from kiara.models.values.data_type import DataTypeClassInfo
info = DataTypeClassInfo.create_from_type_class(self.__class__)
table.add_row("", "")
table.add_row("", Rule())
table.add_row("type_info", info)
return table
characteristics: DataTypeCharacteristics
property
readonly
¶data_type_hash: int
property
readonly
¶data_type_name: str
property
readonly
¶type_config: ~TYPE_CONFIG_CLS
property
readonly
¶Methods¶
assemble_value(self, value_id, data, schema, status, value_hash, pedigree, kiara_id, pedigree_output_name)
¶Source code in kiara/data_types/__init__.py
def assemble_value(
self,
value_id: uuid.UUID,
data: Any,
schema: ValueSchema,
status: Union[ValueStatus, str],
value_hash: int,
pedigree: ValuePedigree,
kiara_id: uuid.UUID,
pedigree_output_name: str,
) -> Tuple[Value, Any]:
if isinstance(status, str):
status = ValueStatus(status).name
this_cls = PythonClass.from_class(self.__class__)
if status in [ValueStatus.SET, ValueStatus.DEFAULT]:
size = self.calculate_size(data)
try:
self._validate(data)
value = Value(
value_id=value_id,
kiara_id=kiara_id,
value_status=status,
value_size=size,
value_hash=value_hash,
value_schema=schema,
pedigree=pedigree,
pedigree_output_name=pedigree_output_name,
data_type_class=this_cls,
)
except Exception as e:
raise KiaraValueException(
data_type=self.__class__, value_data=data, exception=e
)
else:
size = 0
value = Value(
value_id=value_id,
kiara_id=kiara_id,
value_status=status,
value_size=size,
value_hash=value_hash,
value_schema=schema,
pedigree=pedigree,
pedigree_output_name=pedigree_output_name,
data_type_class=this_cls,
)
value._value_data = data
value._data_type = self
return value, data
calculate_hash(self, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/__init__.py
@abc.abstractmethod
def calculate_hash(self, data: TYPE_PYTHON_CLS) -> int:
"""Calculate the hash of the value."""
calculate_size(self, data)
¶Source code in kiara/data_types/__init__.py
@abc.abstractmethod
def calculate_size(self, data: TYPE_PYTHON_CLS) -> int:
pass
create_renderable(self, **config)
¶Source code in kiara/data_types/__init__.py
def create_renderable(self, **config):
show_type_info = config.get("show_type_info", False)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("key")
table.add_column("value", style="i")
table.add_row("type_name", self.data_type_name)
config_json = self.type_config.json(
exclude_unset=True, option=orjson.OPT_INDENT_2
)
config = Syntax(config_json, "json", background_color="default")
table.add_row("type_config", config)
if show_type_info:
from kiara.models.values.data_type import DataTypeClassInfo
info = DataTypeClassInfo.create_from_type_class(self.__class__)
table.add_row("", "")
table.add_row("", Rule())
table.add_row("type_info", info)
return table
data_type_config_class()
classmethod
¶Source code in kiara/data_types/__init__.py
@classmethod
def data_type_config_class(cls) -> Type[TYPE_CONFIG_CLS]:
return DataTypeConfig # type: ignore
parse_python_obj(self, data)
¶Parse a value into a supported python type.
This exists to make it easier to do trivial conversions (e.g. from a date string to a datetime object). If you choose to overwrite this method, make 100% sure that you don't change the meaning of the value, and try to avoid adding or removing information from the data (e.g. by changing the resolution of a date).
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
v |
the value |
required |
Returns:
| Type | Description |
|---|---|
~TYPE_PYTHON_CLS |
'None', if no parsing was done and the original value should be used, otherwise return the parsed Python object |
Source code in kiara/data_types/__init__.py
def parse_python_obj(self, data: Any) -> TYPE_PYTHON_CLS:
"""Parse a value into a supported python type.
This exists to make it easier to do trivial conversions (e.g. from a date string to a datetime object).
If you choose to overwrite this method, make 100% sure that you don't change the meaning of the value, and try to
avoid adding or removing information from the data (e.g. by changing the resolution of a date).
Arguments:
v: the value
Returns:
'None', if no parsing was done and the original value should be used, otherwise return the parsed Python object
"""
return data
python_class()
classmethod
¶Source code in kiara/data_types/__init__.py
@classmethod
@abc.abstractmethod
def python_class(cls) -> Type[TYPE_PYTHON_CLS]:
pass
retrieve_available_type_profiles()
classmethod
¶Source code in kiara/data_types/__init__.py
@classmethod
def retrieve_available_type_profiles(cls) -> Mapping[str, Mapping[str, Any]]:
return {}
DataTypeCharacteristics (BaseModel)
pydantic-model
¶
Source code in kiara/data_types/__init__.py
class DataTypeCharacteristics(BaseModel):
is_scalar: bool = Field(
description="Whether the data desribed by this data type behaves like a skalar.",
default=False,
)
is_json_serializable: bool = Field(
description="Whether the data can be serialized to json without information loss.",
default=False,
)
DataTypeConfig (BaseModel)
pydantic-model
¶
Base class that describes the configuration a [DataType][kiara.data.data_types.DataType] class accepts.
This is stored in the _config_cls class attribute in each DataType class. By default,
a DataType is not configurable, unless the _config_cls class attribute points to a sub-class of this class.
Source code in kiara/data_types/__init__.py
class DataTypeConfig(BaseModel):
"""Base class that describes the configuration a [``DataType``][kiara.data.data_types.DataType] class accepts.
This is stored in the ``_config_cls`` class attribute in each ``DataType`` class. By default,
a ``DataType`` is not configurable, unless the ``_config_cls`` class attribute points to a sub-class of this class.
"""
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
extra = Extra.forbid
@classmethod
def requires_config(cls) -> bool:
"""Return whether this class can be used as-is, or requires configuration before an instance can be created."""
for field_name, field in cls.__fields__.items():
if field.required and field.default is None:
return True
return False
_config_hash: Optional[int] = PrivateAttr(default=None)
def get(self, key: str) -> Any:
"""Get the value for the specified configuation key."""
if key not in self.__fields__:
raise Exception(
f"No config value '{key}' in module config class '{self.__class__.__name__}'."
)
return getattr(self, key)
@property
def config_hash(self) -> int:
if self._config_hash is None:
_d = self.dict()
hashes = DeepHash(_d)
self._config_hash = hashes[_d]
return self._config_hash
def __eq__(self, other):
if self.__class__ != other.__class__:
return False
return self.dict() == other.dict()
def __hash__(self):
return self.config_hash
def __rich_console__(
self, console: Console, options: ConsoleOptions
) -> RenderResult:
my_table = Table(box=box.MINIMAL, show_header=False)
my_table.add_column("Field name", style="i")
my_table.add_column("Value")
for field in self.__fields__:
my_table.add_row(field, getattr(self, field))
yield my_table
config_hash: int
property
readonly
¶
Config
¶Source code in kiara/data_types/__init__.py
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
extra = Extra.forbid
extra
¶json_loads
¶json_dumps(v, *, default=None, **args)
¶Source code in kiara/data_types/__init__.py
def orjson_dumps(v, *, default=None, **args):
# orjson.dumps returns bytes, to match standard json.dumps we need to decode
try:
return orjson.dumps(v, default=default, **args).decode()
except Exception as e:
if is_debug():
print(f"Error dumping json data: {e}")
from kiara import dbg
dbg(v)
raise e
Methods¶
get(self, key)
¶Get the value for the specified configuation key.
Source code in kiara/data_types/__init__.py
def get(self, key: str) -> Any:
"""Get the value for the specified configuation key."""
if key not in self.__fields__:
raise Exception(
f"No config value '{key}' in module config class '{self.__class__.__name__}'."
)
return getattr(self, key)
requires_config()
classmethod
¶Return whether this class can be used as-is, or requires configuration before an instance can be created.
Source code in kiara/data_types/__init__.py
@classmethod
def requires_config(cls) -> bool:
"""Return whether this class can be used as-is, or requires configuration before an instance can be created."""
for field_name, field in cls.__fields__.items():
if field.required and field.default is None:
return True
return False
Modules¶
included_core_types
special
¶
KIARA_MODEL_CLS
¶SCALAR_CHARACTERISTICS
¶Classes¶
AnyType (DataType, Generic)
¶'Any' type, the parent type for most other types.
This type acts as the parents for all (or at least most) non-internal value types. There are some generic operations (like 'persist_value', or 'render_value') which are implemented for this type, so it's descendents have a fallback option in case no subtype-specific operations are implemented for it. In general, it is not recommended to use the 'any' type as module input or output, but it is possible. Values of type 'any' are not allowed to be persisted (at the moment, this might or might not change).
Source code in kiara/data_types/included_core_types/__init__.py
class AnyType(
DataType[TYPE_PYTHON_CLS, DataTypeConfig], Generic[TYPE_PYTHON_CLS, TYPE_CONFIG_CLS]
):
"""'Any' type, the parent type for most other types.
This type acts as the parents for all (or at least most) non-internal value types. There are some generic operations
(like 'persist_value', or 'render_value') which are implemented for this type, so it's descendents have a fallback
option in case no subtype-specific operations are implemented for it. In general, it is not recommended to use the 'any'
type as module input or output, but it is possible. Values of type 'any' are not allowed to be persisted (at the moment,
this might or might not change).
"""
_data_type_name = "any"
@classmethod
def python_class(cls) -> Type:
return object
# def is_immutable(self) -> bool:
# return False
def calculate_hash(self, data: TYPE_PYTHON_CLS) -> int:
return INVALID_HASH_MARKER
# raise Exception(
# f"Calculating the hash for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
def calculate_size(self, data: TYPE_PYTHON_CLS) -> int:
return INVALID_SIZE_MARKER
# raise Exception(
# f"Calculating size for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
def reender_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data = value.data
return str(data)
calculate_hash(self, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/__init__.py
def calculate_hash(self, data: TYPE_PYTHON_CLS) -> int:
return INVALID_HASH_MARKER
# raise Exception(
# f"Calculating the hash for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/__init__.py
def calculate_size(self, data: TYPE_PYTHON_CLS) -> int:
return INVALID_SIZE_MARKER
# raise Exception(
# f"Calculating size for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/__init__.py
@classmethod
def python_class(cls) -> Type:
return object
reender_as__string(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/__init__.py
def reender_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data = value.data
return str(data)
BytesType (AnyType)
¶An array of bytes.
Source code in kiara/data_types/included_core_types/__init__.py
class BytesType(AnyType[bytes, DataTypeConfig]):
"""An array of bytes."""
_data_type_name = "bytes"
@classmethod
def python_class(cls) -> Type:
return bytes
def calculate_hash(cls, data: bytes) -> int:
return KIARA_HASH_FUNCTION(data)
def calculate_size(self, data: bytes) -> int:
return len(data)
def render_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data: bytes = value.data
return data.decode()
calculate_hash(cls, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/__init__.py
def calculate_hash(cls, data: bytes) -> int:
return KIARA_HASH_FUNCTION(data)
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/__init__.py
def calculate_size(self, data: bytes) -> int:
return len(data)
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/__init__.py
@classmethod
def python_class(cls) -> Type:
return bytes
render_as__string(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/__init__.py
def render_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data: bytes = value.data
return data.decode()
KiaraModelValueType (AnyType, Generic)
¶A value type that is used internally.
This type should not be used by user-facing modules and/or operations.
Source code in kiara/data_types/included_core_types/__init__.py
class KiaraModelValueType(
AnyType[KIARA_MODEL_CLS, TYPE_CONFIG_CLS], Generic[KIARA_MODEL_CLS, TYPE_CONFIG_CLS]
):
"""A value type that is used internally.
This type should not be used by user-facing modules and/or operations.
"""
_data_type_name = None # type: ignore
@classmethod
def data_type_config_class(cls) -> Type[DataTypeConfig]:
return DataTypeConfig
@abc.abstractmethod
def create_model_from_python_obj(self, data: Any) -> KIARA_MODEL_CLS:
pass
def parse_python_obj(self, data: Any) -> KIARA_MODEL_CLS:
if isinstance(data, self.__class__.python_class()):
return data # type: ignore
data = self.create_model_from_python_obj(data)
return data
def _validate(self, data: KiaraModel) -> None:
if not isinstance(data, self.__class__.python_class()):
raise Exception(
f"Invalid type '{type(data)}', must be: {self.__class__.python_class().__name__}, or subclass."
)
create_model_from_python_obj(self, data)
¶Source code in kiara/data_types/included_core_types/__init__.py
@abc.abstractmethod
def create_model_from_python_obj(self, data: Any) -> KIARA_MODEL_CLS:
pass
data_type_config_class()
classmethod
¶Source code in kiara/data_types/included_core_types/__init__.py
@classmethod
def data_type_config_class(cls) -> Type[DataTypeConfig]:
return DataTypeConfig
parse_python_obj(self, data)
¶Parse a value into a supported python type.
This exists to make it easier to do trivial conversions (e.g. from a date string to a datetime object). If you choose to overwrite this method, make 100% sure that you don't change the meaning of the value, and try to avoid adding or removing information from the data (e.g. by changing the resolution of a date).
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
v |
the value |
required |
Returns:
| Type | Description |
|---|---|
~KIARA_MODEL_CLS |
'None', if no parsing was done and the original value should be used, otherwise return the parsed Python object |
Source code in kiara/data_types/included_core_types/__init__.py
def parse_python_obj(self, data: Any) -> KIARA_MODEL_CLS:
if isinstance(data, self.__class__.python_class()):
return data # type: ignore
data = self.create_model_from_python_obj(data)
return data
NoneType (DataType)
¶Type indicating a 'None' value
Source code in kiara/data_types/included_core_types/__init__.py
class NoneType(DataType[SpecialValue, DataTypeConfig]):
"""Type indicating a 'None' value"""
_data_type_name = "none"
@classmethod
def python_class(cls) -> Type:
return SpecialValue
# def is_immutable(self) -> bool:
# return False
def calculate_hash(self, data: Any) -> int:
return INVALID_HASH_MARKER
# raise Exception(
# f"Calculating the hash for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
def calculate_size(self, data: Any) -> int:
return INVALID_SIZE_MARKER
# raise Exception(
# f"Calculating size for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
def parse_python_obj(self, data: Any) -> SpecialValue:
return SpecialValue.NO_VALUE
def reender_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data = value.data
return str(data.value)
calculate_hash(self, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/__init__.py
def calculate_hash(self, data: Any) -> int:
return INVALID_HASH_MARKER
# raise Exception(
# f"Calculating the hash for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/__init__.py
def calculate_size(self, data: Any) -> int:
return INVALID_SIZE_MARKER
# raise Exception(
# f"Calculating size for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
parse_python_obj(self, data)
¶Parse a value into a supported python type.
This exists to make it easier to do trivial conversions (e.g. from a date string to a datetime object). If you choose to overwrite this method, make 100% sure that you don't change the meaning of the value, and try to avoid adding or removing information from the data (e.g. by changing the resolution of a date).
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
v |
the value |
required |
Returns:
| Type | Description |
|---|---|
SpecialValue |
'None', if no parsing was done and the original value should be used, otherwise return the parsed Python object |
Source code in kiara/data_types/included_core_types/__init__.py
def parse_python_obj(self, data: Any) -> SpecialValue:
return SpecialValue.NO_VALUE
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/__init__.py
@classmethod
def python_class(cls) -> Type:
return SpecialValue
reender_as__string(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/__init__.py
def reender_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data = value.data
return str(data.value)
StringType (AnyType)
¶A string.
Source code in kiara/data_types/included_core_types/__init__.py
class StringType(AnyType[str, DataTypeConfig]):
"""A string."""
_data_type_name = "string"
@classmethod
def python_class(cls) -> Type:
return str
def is_immutable(self) -> bool:
return True
def calculate_size(self, data: str) -> int:
return len(data)
def calculate_hash(cls, data: str) -> int:
return KIARA_HASH_FUNCTION(data)
def _retrieve_characteristics(self) -> DataTypeCharacteristics:
return SCALAR_CHARACTERISTICS
def validate(cls, value: Any) -> None:
if not isinstance(value, str):
raise ValueError(f"Invalid type '{type(value)}': string required")
def render_as__bytes(self, value: Value, render_config: Mapping[str, Any]):
value_str: str = value.data
return value_str.encode()
calculate_hash(cls, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/__init__.py
def calculate_hash(cls, data: str) -> int:
return KIARA_HASH_FUNCTION(data)
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/__init__.py
def calculate_size(self, data: str) -> int:
return len(data)
is_immutable(self)
¶Source code in kiara/data_types/included_core_types/__init__.py
def is_immutable(self) -> bool:
return True
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/__init__.py
@classmethod
def python_class(cls) -> Type:
return str
render_as__bytes(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/__init__.py
def render_as__bytes(self, value: Value, render_config: Mapping[str, Any]):
value_str: str = value.data
return value_str.encode()
validate(cls, value)
¶Source code in kiara/data_types/included_core_types/__init__.py
def validate(cls, value: Any) -> None:
if not isinstance(value, str):
raise ValueError(f"Invalid type '{type(value)}': string required")
Modules¶
filesystem
¶SUPPORTED_FILE_TYPES
¶logger
¶
FileBundleValueType (AnyType)
¶A bundle of files (like a folder, zip archive, etc.).
Source code in kiara/data_types/included_core_types/filesystem.py
class FileBundleValueType(AnyType[FileBundle, FileTypeConfig]):
"""A bundle of files (like a folder, zip archive, etc.)."""
_data_type_name = "file_bundle"
@classmethod
def retrieve_available_type_profiles(cls) -> Mapping[str, Mapping[str, Any]]:
result = {}
for ft in SUPPORTED_FILE_TYPES:
result[f"{ft}_file_bundle"] = {"content_type": ft}
return result
@classmethod
def python_class(cls) -> Type:
return FileBundle
@classmethod
def data_type_config_class(cls) -> Type[FileTypeConfig]:
return FileTypeConfig
@classmethod
def value_class(cls) -> Type[Value]:
return Value
def calculate_size(self, data: FileBundle) -> int:
return data.size
def calculate_hash(self, data: FileBundle) -> int:
return data.file_bundle_hash
def create_model_from_python_obj(self, data: Any) -> FileBundle:
if isinstance(data, str):
return FileBundle.import_folder(source=data)
else:
raise Exception(
f"Can't create FileBundle from data of type '{type(data)}'."
)
def render_as__terminal_renderable(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
bundle: FileBundle = value.data
renderable = bundle.create_renderable(**render_config)
return renderable
calculate_hash(self, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/filesystem.py
def calculate_hash(self, data: FileBundle) -> int:
return data.file_bundle_hash
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/filesystem.py
def calculate_size(self, data: FileBundle) -> int:
return data.size
create_model_from_python_obj(self, data)
¶Source code in kiara/data_types/included_core_types/filesystem.py
def create_model_from_python_obj(self, data: Any) -> FileBundle:
if isinstance(data, str):
return FileBundle.import_folder(source=data)
else:
raise Exception(
f"Can't create FileBundle from data of type '{type(data)}'."
)
data_type_config_class()
classmethod
¶Source code in kiara/data_types/included_core_types/filesystem.py
@classmethod
def data_type_config_class(cls) -> Type[FileTypeConfig]:
return FileTypeConfig
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/filesystem.py
@classmethod
def python_class(cls) -> Type:
return FileBundle
render_as__terminal_renderable(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/filesystem.py
def render_as__terminal_renderable(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
bundle: FileBundle = value.data
renderable = bundle.create_renderable(**render_config)
return renderable
retrieve_available_type_profiles()
classmethod
¶Source code in kiara/data_types/included_core_types/filesystem.py
@classmethod
def retrieve_available_type_profiles(cls) -> Mapping[str, Mapping[str, Any]]:
result = {}
for ft in SUPPORTED_FILE_TYPES:
result[f"{ft}_file_bundle"] = {"content_type": ft}
return result
value_class()
classmethod
¶Source code in kiara/data_types/included_core_types/filesystem.py
@classmethod
def value_class(cls) -> Type[Value]:
return Value
FileTypeConfig (DataTypeConfig)
pydantic-model
¶
FileValueType (KiaraModelValueType)
¶A file.
Source code in kiara/data_types/included_core_types/filesystem.py
class FileValueType(KiaraModelValueType[FileModel, FileTypeConfig]):
"""A file."""
_data_type_name = "file"
@classmethod
def retrieve_available_type_profiles(cls) -> Mapping[str, Mapping[str, Any]]:
result = {}
for ft in SUPPORTED_FILE_TYPES:
result[f"{ft}_file"] = {"content_type": ft}
return result
@classmethod
def python_class(cls) -> Type:
return FileModel
@classmethod
def data_type_config_class(cls) -> Type[FileTypeConfig]:
return FileTypeConfig
@classmethod
def value_class(cls) -> Type[Value]:
return Value
def calculate_size(self, data: FileModel) -> int:
return data.size
def calculate_hash(self, data: FileModel) -> int:
return KIARA_HASH_FUNCTION(data.file_hash)
def create_model_from_python_obj(self, data: Any) -> FileModel:
if isinstance(data, Mapping):
return FileModel(**data)
if isinstance(data, str):
return FileModel.load_file(source=data)
else:
raise Exception(f"Can't create FileModel from data of type '{type(data)}'.")
def render_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data: Any = value.data
max_lines = render_config.get("max_lines", 34)
try:
lines = []
with open(data.path, "r") as f:
for idx, l in enumerate(f):
if idx > max_lines:
lines.append("...\n")
lines.append("...")
break
lines.append(l)
# TODO: syntax highlighting
return "\n".join(lines)
except UnicodeDecodeError:
# found non-text data
lines = [
"Binary file or non-utf8 enconding, not printing content...",
"",
"[b]File metadata:[/b]",
"",
data.json(option=orjson.OPT_INDENT_2),
]
return "\n".join("lines")
def render_as__terminal_renderable(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data: Any = value.data
max_lines = render_config.get("max_lines", 34)
try:
lines = []
with open(data.path, "r") as f:
for idx, l in enumerate(f):
if idx > max_lines:
lines.append("...\n")
lines.append("...")
break
lines.append(l.rstrip())
return Group(*lines)
except UnicodeDecodeError:
# found non-text data
lines = [
"Binary file or non-utf8 enconding, not printing content...",
"",
"[b]File metadata:[/b]",
"",
data.json(option=orjson.OPT_INDENT_2),
]
return Group(*lines)
calculate_hash(self, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/filesystem.py
def calculate_hash(self, data: FileModel) -> int:
return KIARA_HASH_FUNCTION(data.file_hash)
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/filesystem.py
def calculate_size(self, data: FileModel) -> int:
return data.size
create_model_from_python_obj(self, data)
¶Source code in kiara/data_types/included_core_types/filesystem.py
def create_model_from_python_obj(self, data: Any) -> FileModel:
if isinstance(data, Mapping):
return FileModel(**data)
if isinstance(data, str):
return FileModel.load_file(source=data)
else:
raise Exception(f"Can't create FileModel from data of type '{type(data)}'.")
data_type_config_class()
classmethod
¶Source code in kiara/data_types/included_core_types/filesystem.py
@classmethod
def data_type_config_class(cls) -> Type[FileTypeConfig]:
return FileTypeConfig
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/filesystem.py
@classmethod
def python_class(cls) -> Type:
return FileModel
render_as__string(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/filesystem.py
def render_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data: Any = value.data
max_lines = render_config.get("max_lines", 34)
try:
lines = []
with open(data.path, "r") as f:
for idx, l in enumerate(f):
if idx > max_lines:
lines.append("...\n")
lines.append("...")
break
lines.append(l)
# TODO: syntax highlighting
return "\n".join(lines)
except UnicodeDecodeError:
# found non-text data
lines = [
"Binary file or non-utf8 enconding, not printing content...",
"",
"[b]File metadata:[/b]",
"",
data.json(option=orjson.OPT_INDENT_2),
]
return "\n".join("lines")
render_as__terminal_renderable(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/filesystem.py
def render_as__terminal_renderable(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data: Any = value.data
max_lines = render_config.get("max_lines", 34)
try:
lines = []
with open(data.path, "r") as f:
for idx, l in enumerate(f):
if idx > max_lines:
lines.append("...\n")
lines.append("...")
break
lines.append(l.rstrip())
return Group(*lines)
except UnicodeDecodeError:
# found non-text data
lines = [
"Binary file or non-utf8 enconding, not printing content...",
"",
"[b]File metadata:[/b]",
"",
data.json(option=orjson.OPT_INDENT_2),
]
return Group(*lines)
retrieve_available_type_profiles()
classmethod
¶Source code in kiara/data_types/included_core_types/filesystem.py
@classmethod
def retrieve_available_type_profiles(cls) -> Mapping[str, Mapping[str, Any]]:
result = {}
for ft in SUPPORTED_FILE_TYPES:
result[f"{ft}_file"] = {"content_type": ft}
return result
value_class()
classmethod
¶Source code in kiara/data_types/included_core_types/filesystem.py
@classmethod
def value_class(cls) -> Type[Value]:
return Value
internal
¶
DocumentationModelValueType (InternalModelValueType)
¶Documentation for an internal entity.
Source code in kiara/data_types/included_core_types/internal.py
class DocumentationModelValueType(InternalModelValueType):
"""Documentation for an internal entity."""
_data_type_name = "doc"
def parse_python_obj(self, data: Any) -> DocumentationMetadataModel:
return DocumentationMetadataModel.create(data)
@classmethod
def python_class(cls) -> Type:
return DocumentationMetadataModel
def render_as__terminal_renderable(
self, value: "Value", render_config: Mapping[str, Any]
):
json_str = value.data.json(option=orjson.OPT_INDENT_2)
return Syntax(json_str, "json", background_color="default")
parse_python_obj(self, data)
¶Parse a value into a supported python type.
This exists to make it easier to do trivial conversions (e.g. from a date string to a datetime object). If you choose to overwrite this method, make 100% sure that you don't change the meaning of the value, and try to avoid adding or removing information from the data (e.g. by changing the resolution of a date).
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
v |
the value |
required |
Returns:
| Type | Description |
|---|---|
DocumentationMetadataModel |
'None', if no parsing was done and the original value should be used, otherwise return the parsed Python object |
Source code in kiara/data_types/included_core_types/internal.py
def parse_python_obj(self, data: Any) -> DocumentationMetadataModel:
return DocumentationMetadataModel.create(data)
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/internal.py
@classmethod
def python_class(cls) -> Type:
return DocumentationMetadataModel
render_as__terminal_renderable(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/internal.py
def render_as__terminal_renderable(
self, value: "Value", render_config: Mapping[str, Any]
):
json_str = value.data.json(option=orjson.OPT_INDENT_2)
return Syntax(json_str, "json", background_color="default")
InternalModelValueType (InternalType)
¶A value type that is used internally.
This type should not be used by user-facing modules and/or operations.
Source code in kiara/data_types/included_core_types/internal.py
class InternalModelValueType(InternalType[KiaraModel, DataTypeConfig]):
"""A value type that is used internally.
This type should not be used by user-facing modules and/or operations.
"""
_data_type_name = "internal_model"
@classmethod
def python_class(cls) -> Type:
return KiaraModel
@classmethod
def data_type_config_class(cls) -> Type[DataTypeConfig]:
return DataTypeConfig
def calculate_size(self, data: KiaraModel) -> int:
return data.model_size
def calculate_hash(self, data: KiaraModel) -> int:
return data.model_data_hash
def _validate(self, value: KiaraModel) -> None:
if not isinstance(value, KiaraModel):
raise Exception(f"Invalid type: {type(value)}.")
def render_as__terminal_renderable(
self, value: "Value", render_config: Mapping[str, Any]
):
json_str = value.data.json(option=orjson.OPT_INDENT_2)
return Syntax(json_str, "json", background_color="default")
calculate_hash(self, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/internal.py
def calculate_hash(self, data: KiaraModel) -> int:
return data.model_data_hash
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/internal.py
def calculate_size(self, data: KiaraModel) -> int:
return data.model_size
data_type_config_class()
classmethod
¶Source code in kiara/data_types/included_core_types/internal.py
@classmethod
def data_type_config_class(cls) -> Type[DataTypeConfig]:
return DataTypeConfig
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/internal.py
@classmethod
def python_class(cls) -> Type:
return KiaraModel
render_as__terminal_renderable(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/internal.py
def render_as__terminal_renderable(
self, value: "Value", render_config: Mapping[str, Any]
):
json_str = value.data.json(option=orjson.OPT_INDENT_2)
return Syntax(json_str, "json", background_color="default")
InternalType (DataType, Generic)
¶'A 'marker' base data type for data types that are (mainly) used internally in kiara..
Source code in kiara/data_types/included_core_types/internal.py
class InternalType(
DataType[TYPE_PYTHON_CLS, TYPE_CONFIG_CLS],
Generic[TYPE_PYTHON_CLS, TYPE_CONFIG_CLS],
):
"""'A 'marker' base data type for data types that are (mainly) used internally in kiara.."""
_data_type_name = "internal"
@classmethod
def python_class(cls) -> Type:
return object
# def is_immutable(self) -> bool:
# return False
def calculate_hash(self, data: TYPE_PYTHON_CLS) -> int:
return INVALID_HASH_MARKER
# raise Exception(
# f"Calculating the hash for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
def calculate_size(self, data: TYPE_PYTHON_CLS) -> int:
return INVALID_SIZE_MARKER
# raise Exception(
# f"Calculating size for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
def reender_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data = value.data
return str(data)
calculate_hash(self, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/internal.py
def calculate_hash(self, data: TYPE_PYTHON_CLS) -> int:
return INVALID_HASH_MARKER
# raise Exception(
# f"Calculating the hash for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/internal.py
def calculate_size(self, data: TYPE_PYTHON_CLS) -> int:
return INVALID_SIZE_MARKER
# raise Exception(
# f"Calculating size for type '{self.__class__._value_type_name}' is not supported. If your type inherits from 'any', make sure to implement the 'calculate_hash' method."
# )
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/internal.py
@classmethod
def python_class(cls) -> Type:
return object
reender_as__string(self, value, render_config)
¶Source code in kiara/data_types/included_core_types/internal.py
def reender_as__string(
self, value: "Value", render_config: Mapping[str, Any]
) -> Any:
data = value.data
return str(data)
TerminalRenderable (InternalType)
¶A list of renderable objects, used in the 'rich' Python library, to print to the terminal or in Jupyter.
Internally, the result list items can be either a string, a 'rich.console.ConsoleRenderable', or a 'rich.console.RichCast'.
Source code in kiara/data_types/included_core_types/internal.py
class TerminalRenderable(InternalType[object, DataTypeConfig]):
"""A list of renderable objects, used in the 'rich' Python library, to print to the terminal or in Jupyter.
Internally, the result list items can be either a string, a 'rich.console.ConsoleRenderable', or a 'rich.console.RichCast'.
"""
_value_type_name = "terminal_renderable"
@classmethod
def python_class(cls) -> Type:
return object
def calculate_hash(self, data: TYPE_PYTHON_CLS) -> int:
return compute_hash(data)
def calculate_size(self, data: TYPE_PYTHON_CLS) -> int:
return sys.getsizeof(data)
calculate_hash(self, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/internal.py
def calculate_hash(self, data: TYPE_PYTHON_CLS) -> int:
return compute_hash(data)
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/internal.py
def calculate_size(self, data: TYPE_PYTHON_CLS) -> int:
return sys.getsizeof(data)
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/internal.py
@classmethod
def python_class(cls) -> Type:
return object
persistence
¶
LoadConfigSchema (DataTypeConfig)
pydantic-model
¶Source code in kiara/data_types/included_core_types/persistence.py
class LoadConfigSchema(DataTypeConfig):
persistence_target: str = Field(
description="A hint as to the persistence target (e.g. disk)."
)
persistence_format: str = Field(
description="A hint as to the persistence format (e.g. pickle)."
)
LoadConfigValueType (InternalType)
¶A value type that contains data that describes how to (re-)load a value from disk.
This is mostly used internally in kiara, but might be exposed to users in certain cases (for example when exporting a value in the native kiara file format).
Source code in kiara/data_types/included_core_types/persistence.py
class LoadConfigValueType(InternalType[LoadConfig, LoadConfigSchema]):
"""A value type that contains data that describes how to (re-)load a value from disk.
This is mostly used internally in kiara, but might be exposed to users in certain cases (for example when exporting
a value in the native kiara file format).
"""
_data_type_name = "load_config"
@classmethod
def python_class(cls) -> Type:
return LoadConfig
@classmethod
def data_type_config_class(cls) -> Type[LoadConfigSchema]:
return LoadConfigSchema
def is_immutable(self) -> bool:
return True
def calculate_hash(self, data: LoadConfig) -> int:
"""Calculate the hash of the value."""
return data.model_data_hash
def calculate_size(self, data: LoadConfig) -> int:
return data.model_size
@property
def persistence_target(self) -> str:
return self.type_config.persistence_target
@property
def persistence_format(self) -> str:
return self.type_config.persistence_format
persistence_format: str
property
readonly
¶persistence_target: str
property
readonly
¶calculate_hash(self, data)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/persistence.py
def calculate_hash(self, data: LoadConfig) -> int:
"""Calculate the hash of the value."""
return data.model_data_hash
calculate_size(self, data)
¶Source code in kiara/data_types/included_core_types/persistence.py
def calculate_size(self, data: LoadConfig) -> int:
return data.model_size
data_type_config_class()
classmethod
¶Source code in kiara/data_types/included_core_types/persistence.py
@classmethod
def data_type_config_class(cls) -> Type[LoadConfigSchema]:
return LoadConfigSchema
is_immutable(self)
¶Source code in kiara/data_types/included_core_types/persistence.py
def is_immutable(self) -> bool:
return True
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/persistence.py
@classmethod
def python_class(cls) -> Type:
return LoadConfig
serialization
¶
DeserializationConfig (Manifest)
pydantic-model
¶Source code in kiara/data_types/included_core_types/serialization.py
class DeserializationConfig(Manifest):
output_name: str = Field(
description="The name of the field that contains the deserialized value."
)
SerializedValueModel (BaseModel)
pydantic-model
¶Source code in kiara/data_types/included_core_types/serialization.py
class SerializedValueModel(BaseModel):
deserialization_config: DeserializationConfig = Field(
description="The configuration for a kiara module that deserializes this value."
)
data: Dict[str, bytes] = Field(
description="One or several byte arrays representing the serialized state of the value."
)
_cached_size: Optional[int] = PrivateAttr(default=None)
_cached_hash: Optional[int] = PrivateAttr(default=None)
@property
def serialized_size(self) -> int:
if self._cached_size is not None:
return self._cached_size
size = 0
for k, v in self.data.items():
size = size + len(k) + len(v)
self._cached_size = size
return self._cached_size
@property
def serialized_hash(self) -> int:
if self._cached_hash is not None:
return self._cached_hash
obj = {
"deserialization_config": self.deserialization_config.dict(),
"data": {k: hash_from_buffer(v) for k, v in self.data.items()},
}
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
self._cached_hash = h[obj]
return self._cached_hash
def __repr__(self):
return f"{self.__class__.__name__}(deserialization_config={self.deserialization_config}, size={self.serialized_size}, hash={self.serialized_hash})"
def __str__(self):
return self.__repr__()
data: Dict[str, bytes]
pydantic-field
required
¶One or several byte arrays representing the serialized state of the value.
deserialization_config: DeserializationConfig
pydantic-field
required
¶The configuration for a kiara module that deserializes this value.
serialized_hash: int
property
readonly
¶serialized_size: int
property
readonly
¶
SerializedValueType (InternalType)
¶A data type that contains a serialized representation of a value.
This is used for transferring/streaming value over the wire, and works on a similar principle as the 'load_config' value type.
Source code in kiara/data_types/included_core_types/serialization.py
class SerializedValueType(
InternalType[SerializedValueModel, SerializedValueTypeConfigSchema]
):
"""A data type that contains a serialized representation of a value.
This is used for transferring/streaming value over the wire, and works on a similar principle as the 'load_config'
value type.
"""
@classmethod
def python_class(cls) -> Type:
return SerializedValueModel
@classmethod
def data_type_config_class(cls) -> Type[SerializedValueTypeConfigSchema]:
return SerializedValueTypeConfigSchema
def is_immutable(self) -> bool:
return True
def calculate_hash(self, value: SerializedValueModel) -> int:
"""Calculate the hash of the value."""
return value.serialized_hash
def calculate_size(self, value: SerializedValueModel) -> int:
return value.serialized_size
@property
def format_name(self) -> str:
return self.type_config.format_name
format_name: str
property
readonly
¶calculate_hash(self, value)
¶Calculate the hash of the value.
Source code in kiara/data_types/included_core_types/serialization.py
def calculate_hash(self, value: SerializedValueModel) -> int:
"""Calculate the hash of the value."""
return value.serialized_hash
calculate_size(self, value)
¶Source code in kiara/data_types/included_core_types/serialization.py
def calculate_size(self, value: SerializedValueModel) -> int:
return value.serialized_size
data_type_config_class()
classmethod
¶Source code in kiara/data_types/included_core_types/serialization.py
@classmethod
def data_type_config_class(cls) -> Type[SerializedValueTypeConfigSchema]:
return SerializedValueTypeConfigSchema
is_immutable(self)
¶Source code in kiara/data_types/included_core_types/serialization.py
def is_immutable(self) -> bool:
return True
python_class()
classmethod
¶Source code in kiara/data_types/included_core_types/serialization.py
@classmethod
def python_class(cls) -> Type:
return SerializedValueModel
SerializedValueTypeConfigSchema (DataTypeConfig)
pydantic-model
¶
defaults
¶
Attributes¶
ANY_TYPE_NAME
¶
ARRAY_MODEL_CATEOGORY_ID
¶
AUTHORS_METADATA_CATEGORY_ID
¶
BATCH_CONFIG_TYPE_CATEGORY_ID
¶
COLOR_LIST
¶
CONTEXT_INFO_CATEGORY_ID
¶
CONTEXT_METADATA_CATEOGORY_ID
¶
DATA_TYPES_CATEGORY_ID
¶
DATA_TYPE_CATEGORY_ID
¶
DATA_TYPE_CLASS_CATEGORY_ID
¶
DATA_WRAP_CATEGORY_ID
¶
DEFAULT_ALIAS_STORE_MARKER
¶
Name for the default context job store.
DEFAULT_CONTEXT_NAME
¶
DEFAULT_DATA_STORE_MARKER
¶
Name for the default context data store.
DEFAULT_EXCLUDE_DIRS
¶
List of directory names to exclude by default when walking a folder recursively.
DEFAULT_EXCLUDE_FILES
¶
List of file names to exclude by default when reading folders.
DEFAULT_JOB_STORE_MARKER
¶
Name for the default context job store.
DEFAULT_NO_DESC_VALUE
¶
DEFAULT_PIPELINE_PARENT_ID
¶
Default parent id for pipeline objects that are not associated with a workflow.
DEFAULT_PRETTY_PRINT_CONFIG
¶
DEFAULT_TO_JSON_CONFIG: Mapping[str, Any]
¶
DESTINY_CATEGORY_ID
¶
DOCUMENTATION_CATEGORY_ID
¶
ENVIRONMENT_TYPE_CATEGORY_ID
¶
FILE_BUNDLE_MODEL_CATEOGORY_ID
¶
FILE_MODEL_CATEOGORY_ID
¶
INVALID_HASH_MARKER
¶
INVALID_SIZE_MARKER
¶
INVALID_VALUE_NAMES
¶
List of reserved names, inputs/outputs can't use those.
JOB_CATEGORY_ID
¶
JOB_CONFIG_TYPE_CATEGORY_ID
¶
JOB_LOG_CATEGORY_ID
¶
JOB_RECORD_TYPE_CATEGORY_ID
¶
KIARA_ALIASES_DIR
¶
KIARA_DATA_DIR
¶
KIARA_DATA_STORE_DIR
¶
KIARA_DB_FILE
¶
KIARA_DB_MIGRATIONS_CONFIG
¶
KIARA_DB_MIGRATIONS_FOLDER
¶
KIARA_DEFAULT_ROOT_NODE_ID
¶
KIARA_HASH_FUNCTION
¶
KIARA_MAIN_CONFIG_FILE
¶
KIARA_METADATA_DIR
¶
KIARA_MODULE_BASE_FOLDER
¶
Marker to indicate the base folder for the kiara module.
KIARA_MODULE_METADATA_ATTRIBUTE
¶
KIARA_RESOURCES_FOLDER
¶
Default resources folder for this package.
KIARA_ROOT_TYPE_NAME
¶
LOAD_CONFIG_DATA_TYPE_NAME
¶
LOAD_CONFIG_PLACEHOLDER
¶
MODULE_CONFIG_CATEGORY_ID
¶
MODULE_CONFIG_METADATA_CATEGORY_ID
¶
MODULE_CONFIG_SCHEMA_CATEGORY_ID
¶
MODULE_TYPES_CATEGORY_ID
¶
MODULE_TYPE_CATEGORY_ID
¶
MODULE_TYPE_KEY
¶
The key to specify the type of a module.
MODULE_TYPE_NAME_KEY
¶
The string for the module type name in a module configuration dict.
NONE_VALUE_ID
¶
NOT_SET_VALUE_ID
¶
NO_HASH_MARKER
¶
Marker string to indicate no hash was calculated.
NO_MODULE_TYPE
¶
NO_VALUE_ID_MARKER
¶
Marker string to indicate no value id exists.
OPERATIONS_CATEGORY_ID
¶
OPERATION_CATEOGORY_ID
¶
OPERATION_CONFIG_CATEOGORY_ID
¶
OPERATION_DETAILS_CATEOGORY_ID
¶
OPERATION_INPUTS_SCHEMA_CATEOGORY_ID
¶
OPERATION_OUTPUTS_SCHEMA_CATEOGORY_ID
¶
OPERATION_TYPES_CATEGORY_ID
¶
OPERATION_TYPE_CATEGORY_ID
¶
ORPHAN_PEDIGREE_OUTPUT_NAME
¶
PIPELINE_CONFIG_TYPE_CATEGORY_ID
¶
PIPELINE_PARENT_MARKER
¶
Marker string in the pipeline structure that indicates a parent pipeline element.
PIPELINE_STEP_DETAILS_CATEGORY_ID
¶
PIPELINE_STEP_TYPE_CATEGORY_ID
¶
PIPELINE_STRUCTURE_TYPE_CATEGORY_ID
¶
PIPELINE_TYPES_CATEGORY_ID
¶
PIPELINE_TYPE_CATEGORY_ID
¶
PYDANTIC_USE_CONSTRUCT: bool
¶
SERIALIZED_DATA_TYPE_NAME
¶
STEP_ID_KEY
¶
The key to specify the step id.
STRICT_CHECKS: bool
¶
TABLE_MODEL_CATEOGORY_ID
¶
UNOLOADABLE_DATA_CATEGORY_ID
¶
USER_PIPELINES_FOLDER
¶
VALID_PIPELINE_FILE_EXTENSIONS
¶
File extensions a kiara pipeline/workflow file can have.
VALUES_CATEGORY_ID
¶
VALUE_CATEGORY_ID
¶
VALUE_METADATA_CATEGORY_ID
¶
VALUE_PEDIGREE_TYPE_CATEGORY_ID
¶
VALUE_SCHEMA_CATEGORY_ID
¶
VOID_KIARA_ID
¶
kiara_app_dirs
¶
Classes¶
doc
special
¶
Main module for code that helps with documentation auto-generation in supported projects.
Classes¶
FrklDocumentationPlugin (BasePlugin)
¶
mkdocs plugin to render API documentation for a project.
To add to a project, add this to the 'plugins' section of a mkdocs config file:
- frkl-docgen:
main_module: "module_name"
This will add an API reference navigation item to your page navigation, with auto-generated entries for every
Python module in your package.
Source code in kiara/doc/__init__.py
class FrklDocumentationPlugin(BasePlugin):
"""[mkdocs](https://www.mkdocs.org/) plugin to render API documentation for a project.
To add to a project, add this to the 'plugins' section of a mkdocs config file:
```yaml
- frkl-docgen:
main_module: "module_name"
```
This will add an ``API reference`` navigation item to your page navigation, with auto-generated entries for every
Python module in your package.
"""
config_scheme = (("main_module", mkdocs.config.config_options.Type(str)),)
def __init__(self):
self._doc_paths = None
self._dir = tempfile.TemporaryDirectory(prefix="frkl_doc_gen_")
self._doc_files = None
super().__init__()
def on_files(self, files: Files, config: Config) -> Files:
self._doc_paths = gen_pages_for_module(self.config["main_module"])
self._doc_files = {}
for k in sorted(self._doc_paths, key=lambda x: os.path.splitext(x)[0]):
content = self._doc_paths[k]["content"]
_file = File(
k,
src_dir=self._dir.name,
dest_dir=config["site_dir"],
use_directory_urls=config["use_directory_urls"],
)
os.makedirs(os.path.dirname(_file.abs_src_path), exist_ok=True)
with open(_file.abs_src_path, "w") as f:
f.write(content)
self._doc_files[k] = _file
files.append(_file)
return files
def on_page_content(self, html, page: Page, config: Config, files: Files):
repo_url = config.get("repo_url", None)
python_src = config.get("edit_uri", None)
if page.file.src_path in self._doc_paths.keys():
src_path = self._doc_paths.get(page.file.src_path)["python_src"]["rel_path"]
rel_base = urllib.parse.urljoin(repo_url, f"{python_src}/../src/{src_path}")
page.edit_url = rel_base
return html
def on_nav(self, nav: Navigation, config: Config, files: Files):
for item in nav.items:
if item.title and "Api reference" in item.title:
return nav
pages = []
for _file in self._doc_files.values():
pages.append(_file.page)
section = Section(title="API reference", children=pages)
nav.items.append(section)
nav.pages.extend(pages)
_add_previous_and_next_links(nav.pages)
_add_parent_links(nav.items)
return nav
def on_post_build(self, config: Config):
self._dir.cleanup()
config_scheme
¶on_files(self, files, config)
¶Source code in kiara/doc/__init__.py
def on_files(self, files: Files, config: Config) -> Files:
self._doc_paths = gen_pages_for_module(self.config["main_module"])
self._doc_files = {}
for k in sorted(self._doc_paths, key=lambda x: os.path.splitext(x)[0]):
content = self._doc_paths[k]["content"]
_file = File(
k,
src_dir=self._dir.name,
dest_dir=config["site_dir"],
use_directory_urls=config["use_directory_urls"],
)
os.makedirs(os.path.dirname(_file.abs_src_path), exist_ok=True)
with open(_file.abs_src_path, "w") as f:
f.write(content)
self._doc_files[k] = _file
files.append(_file)
return files
on_nav(self, nav, config, files)
¶Source code in kiara/doc/__init__.py
def on_nav(self, nav: Navigation, config: Config, files: Files):
for item in nav.items:
if item.title and "Api reference" in item.title:
return nav
pages = []
for _file in self._doc_files.values():
pages.append(_file.page)
section = Section(title="API reference", children=pages)
nav.items.append(section)
nav.pages.extend(pages)
_add_previous_and_next_links(nav.pages)
_add_parent_links(nav.items)
return nav
on_page_content(self, html, page, config, files)
¶Source code in kiara/doc/__init__.py
def on_page_content(self, html, page: Page, config: Config, files: Files):
repo_url = config.get("repo_url", None)
python_src = config.get("edit_uri", None)
if page.file.src_path in self._doc_paths.keys():
src_path = self._doc_paths.get(page.file.src_path)["python_src"]["rel_path"]
rel_base = urllib.parse.urljoin(repo_url, f"{python_src}/../src/{src_path}")
page.edit_url = rel_base
return html
on_post_build(self, config)
¶Source code in kiara/doc/__init__.py
def on_post_build(self, config: Config):
self._dir.cleanup()
Modules¶
gen_info_pages
¶
generate_detail_pages(context_info, sub_path='info', add_summary_page=False)
¶Source code in kiara/doc/gen_info_pages.py
def generate_detail_pages(
context_info: KiaraContextInfo,
sub_path: str = "info",
add_summary_page: bool = False,
):
pages = {}
summary = []
all_info = context_info.get_all_info(skip_empty_types=True)
for item_type, items_info in all_info.items():
summary.append(f"* [{item_type}]({item_type}.md)")
path = render_item_listing(
item_type=item_type, items=items_info, sub_path=sub_path
)
pages[item_type] = path
if summary:
if add_summary_page:
summary.insert(0, "* [Summary](index.md)")
with mkdocs_gen_files.open(f"{sub_path}/SUMMARY.md", "w") as f:
f.write("\n".join(summary))
return pages
get_jina_env()
¶Source code in kiara/doc/gen_info_pages.py
def get_jina_env():
global _jinja_env
if _jinja_env is None:
from jinja2 import Environment, FileSystemLoader
_jinja_env = Environment(
loader=FileSystemLoader(
os.path.join(KIARA_RESOURCES_FOLDER, "templates", "doc_gen"),
encoding="utf8",
)
)
return _jinja_env
render_item_listing(item_type, items, sub_path='info')
¶Source code in kiara/doc/gen_info_pages.py
def render_item_listing(item_type: str, items: InfoModelGroup, sub_path: str = "info"):
list_template = get_jina_env().get_template("info_listing.j2")
render_args = {"items": items.get_type_infos(), "item_type": item_type}
rendered = list_template.render(**render_args)
path = f"{sub_path}/{item_type}.md"
with mkdocs_gen_files.open(path, "w") as f:
f.write(rendered)
return path
generate_api_doc
¶
Functions¶
gen_pages_for_module(module, prefix='api_reference')
¶Generate modules for a set of modules (using the mkdocstring package.
Source code in kiara/doc/generate_api_doc.py
def gen_pages_for_module(
module: typing.Union[str, ModuleType], prefix: str = "api_reference"
):
"""Generate modules for a set of modules (using the [mkdocstring](https://github.com/mkdocstrings/mkdocstrings) package."""
result = {}
modules_info = get_source_tree(module)
for module_name, path in modules_info.items():
page_name = module_name
if page_name.endswith("__init__"):
page_name = page_name[0:-9]
if page_name.endswith("._frkl"):
continue
doc_path = f"{prefix}{os.path.sep}{page_name}.md"
p = Path("..", path["abs_path"])
if not p.read_text().strip():
continue
main_module = path["main_module"]
if page_name == main_module:
title = page_name
else:
title = page_name.replace(f"{main_module}.", "➜ ")
result[doc_path] = {
"python_src": path,
"content": f"---\ntitle: {title}\n---\n# {page_name}\n\n::: {module_name}",
}
return result
get_source_tree(module)
¶Find all python source files for a module.
Source code in kiara/doc/generate_api_doc.py
def get_source_tree(module: typing.Union[str, ModuleType]):
"""Find all python source files for a module."""
if isinstance(module, str):
module = importlib.import_module(module)
if not isinstance(module, ModuleType):
raise TypeError(
f"Invalid type '{type(module)}', input needs to be a string or module."
)
module_file = module.__file__
assert module_file is not None
module_root = os.path.dirname(module_file)
module_name = module.__name__
src = {}
for path in Path(module_root).glob("**/*.py"):
rel = os.path.relpath(path, module_root)
mod_name = f"{module_name}.{rel[0:-3].replace(os.path.sep, '.')}"
rel_path = f"{module_name}{os.path.sep}{rel}"
src[mod_name] = {
"rel_path": rel_path,
"abs_path": path,
"main_module": module_name,
}
return src
mkdocs_macros_cli
¶
CACHE_DIR
¶os_env_vars
¶Functions¶
define_env(env)
¶Helper macros for Python project documentation.
Currently, those macros are available (check the source code for more details):
cli¶
Execute a command on the command-line, capture the output and return it to be used in a documentation page.
inline_file_as_codeblock¶
Read an external file, and return its content as a markdown code block.
Source code in kiara/doc/mkdocs_macros_cli.py
def define_env(env):
"""
Helper macros for Python project documentation.
Currently, those macros are available (check the source code for more details):
## ``cli``
Execute a command on the command-line, capture the output and return it to be used in a documentation page.
## ``inline_file_as_codeblock``
Read an external file, and return its content as a markdown code block.
"""
# env.variables["baz"] = "John Doe"
@env.macro
def cli(
*command,
print_command: bool = True,
code_block: bool = True,
split_command_and_output: bool = True,
max_height: Optional[int] = None,
cache_key: Optional[str] = None,
extra_env: Optional[Dict[str, str]] = None,
fake_command: Optional[str] = None,
):
"""Execute the provided command, save the output and return it to be used in documentation modules."""
hashes = DeepHash(command)
hash_str = hashes[command]
hashes_env = DeepHash(extra_env)
hashes_env_str = hashes_env[extra_env]
hash_str = hash_str + "_" + hashes_env_str
if cache_key:
hash_str = hash_str + "_" + cache_key
cache_file: Path = Path(os.path.join(CACHE_DIR, str(hash_str)))
_run_env = dict(os_env_vars)
if extra_env:
_run_env.update(extra_env)
if cache_file.is_file():
stdout = cache_file.read_text()
else:
try:
print(f"RUNNING: {' '.join(command)}")
result = subprocess.check_output(command, env=_run_env)
stdout = result.decode()
cache_file.write_text(stdout)
except subprocess.CalledProcessError as e:
stdout = f"Error: {e}\n\nStdout: {e.stdout}\n\nStderr: {e.stderr}"
print("stdout:")
print(e.stdout)
print("stderr:")
print(e.stderr)
if os.getenv("FAIL_DOC_BUILD_ON_ERROR") == "true":
sys.exit(1)
if fake_command:
command_str = fake_command
else:
command_str = " ".join(command)
if split_command_and_output and print_command:
_c = f"\n``` console\n{command_str}\n```\n"
_output = "``` console\n" + stdout + "\n```\n"
if max_height is not None and max_height > 0:
_output = f"<div style='max-height:{max_height}px;overflow:auto'>\n{_output}\n</div>"
_stdout = _c + _output
else:
if print_command:
_stdout = f"> {command_str}\n{stdout}"
if code_block:
_stdout = "``` console\n" + _stdout + "\n```\n"
if max_height is not None and max_height > 0:
_stdout = f"<div style='max-height:{max_height}px;overflow:auto'>\n{_stdout}\n</div>"
return _stdout
@env.macro
def inline_file_as_codeblock(path, format: str = ""):
"""Import external file and return its content as a markdown code block."""
f = Path(path)
return f"```{format}\n{f.read_text()}\n```"
mkdocs_macros_kiara
¶
kiara_obj
¶yaml
¶Functions¶
define_env(env)
¶This is the hook for defining variables, macros and filters
- variables: the dictionary that contains the environment variables
- macro: a decorator function, to declare a macro.
Source code in kiara/doc/mkdocs_macros_kiara.py
def define_env(env):
"""
This is the hook for defining variables, macros and filters
- variables: the dictionary that contains the environment variables
- macro: a decorator function, to declare a macro.
"""
# env.variables["baz"] = "John Doe"
@env.macro
def get_schema_for_model(model_class: Union[str, Type[BaseModel]]):
if isinstance(model_class, str):
_class: Type[BaseModel] = locate(model_class) # type: ignore
else:
_class = model_class
schema_json = _class.schema_json(indent=2)
return schema_json
@env.macro
def get_src_of_object(obj: Union[str, Any]):
try:
if isinstance(obj, str):
_obj: Type[BaseModel] = locate(obj) # type: ignore
else:
_obj = obj
src = inspect.getsource(_obj)
return src
except Exception as e:
return f"Can't render object source: {str(e)}"
@env.macro
def get_context_info() -> KiaraContextInfo:
return builtins.plugin_package_context_info # type: ignore
# @env.macro
# def get_module_info(module_type: str):
#
# try:
#
# m_cls = Kiara.instance().module_registry.get_module_class(module_type)
# info = KiaraModuleTypeInfo.from_module_class(m_cls)
#
# from rich.console import Console
#
# console = Console(record=True)
# console.print(info)
#
# html = console.export_text()
# return html
# except Exception as e:
# return f"Can't render module info: {str(e)}"
#
# @env.macro
# def get_info_item_list_for_category(
# category: str, limit_to_package: typing.Optional[str] = None
# ) -> typing.Dict[str, KiaraInfoModel]:
# return _get_info_item_list_for_category(
# category=category, limit_to_package=limit_to_package
# )
#
# def _get_info_item_list_for_category(
# category: str, limit_to_package: typing.Optional[str] = None
# ) -> typing.Dict[str, KiaraInfoModel]:
#
# infos = kiara_context.find_subcomponents(category=category)
#
# if limit_to_package:
# temp = {}
# for n_id, obj in infos.items():
# if obj.context.labels.get("package", None) == limit_to_package:
# temp[n_id] = obj
# infos = temp
#
# docs = {}
# for n_id, obj in infos.items():
# docs[obj.get_id()] = obj.documentation.description
#
# return docs
#
# @env.macro
# def get_info_for_categories(
# *categories: str, limit_to_package: typing.Optional[str] = None
# ):
#
# TITLE_MAP = {
# "metadata.module": "Modules",
# "metadata.pipeline": "Pipelines",
# "metadata.type": "Value data_types",
# "metadata.operation_type": "Operation data_types",
# }
# result = {}
# for cat in categories:
# infos = _get_info_item_list_for_category(
# cat, limit_to_package=limit_to_package
# )
# if infos:
# result[cat] = {"items": infos, "title": TITLE_MAP[cat]}
#
# return result
#
# @env.macro
# def get_module_list_for_package(
# package_name: str,
# include_core_modules: bool = True,
# include_pipelines: bool = True,
# ):
#
# modules = kiara_obj.module_registry.find_modules_for_package(
# package_name,
# include_core_modules=include_core_modules,
# include_pipelines=include_pipelines,
# )
#
# result = []
# for name, info in modules.items():
# type_md = info.get_type_metadata()
# result.append(
# f"[``{name}``][kiara_info.modules.{name}]: {type_md.documentation.description}"
# )
#
# return result
#
# @env.macro
# def get_data_types_for_package(package_name: str):
#
# data_types = kiara_obj.type_registry.find_data_type_classes_for_package(
# package_name
# )
# result = []
# for name, info in data_types.items():
# type_md = info.get_type_metadata()
# result.append(f" - ``{name}``: {type_md.documentation.description}")
#
# return "\n".join(result)
#
# @env.macro
# def get_metadata_models_for_package(package_name: str):
#
# metadata_schemas = kiara_obj.metadata_mgmt.find_all_models_for_package(
# package_name
# )
# result = []
# for name, info in metadata_schemas.items():
# type_md = info.get_type_metadata()
# result.append(f" - ``{name}``: {type_md.documentation.description}")
#
# return "\n".join(result)
#
# @env.macro
# def get_kiara_context() -> KiaraContext:
# return kiara_context
mkdocstrings
special
¶
Modules¶
collector
¶logger
¶
KiaraCollector (BaseCollector)
¶The class responsible for loading Jinja templates and rendering them.
It defines some configuration options, implements the render method,
and overrides the update_env method of the [BaseRenderer class][mkdocstrings.handlers.base.BaseRenderer].
Source code in kiara/doc/mkdocstrings/collector.py
class KiaraCollector(BaseCollector):
"""The class responsible for loading Jinja templates and rendering them.
It defines some configuration options, implements the `render` method,
and overrides the `update_env` method of the [`BaseRenderer` class][mkdocstrings.handlers.base.BaseRenderer].
"""
default_config: dict = {"docstring_style": "google", "docstring_options": {}}
"""The default selection options.
Option | Type | Description | Default
------ | ---- | ----------- | -------
**`docstring_style`** | `"google" | "numpy" | "sphinx" | None` | The docstring style to use. | `"google"`
**`docstring_options`** | `dict[str, Any]` | The options for the docstring parser. | `{}`
"""
fallback_config: dict = {"fallback": True}
def __init__(self) -> None:
"""Initialize the collector."""
self._kiara: Kiara = Kiara.instance()
def collect(self, identifier: str, config: dict) -> CollectorItem: # noqa: WPS231
"""Collect the documentation tree given an identifier and selection options.
Arguments:
identifier: The dotted-path of a Python object available in the Python path.
config: Selection options, used to alter the data collection done by `pytkdocs`.
Raises:
CollectionError: When there was a problem collecting the object documentation.
Returns:
The collected object-tree.
"""
tokens = identifier.split(".")
if tokens[0] != "kiara_info":
return None
item_type = tokens[1]
item_id = ".".join(tokens[2:])
ctx: KiaraContextInfo = builtins.plugin_package_context_info # type: ignore
try:
item: KiaraInfoModel = ctx.get_info(item_type=item_type, item_id=item_id)
except Exception:
import traceback
traceback.print_exc()
raise CollectionError(f"Invalid id: {identifier}")
return {"obj": item, "identifier": identifier}
default_config: dict
¶The default selection options.
Option | Type | Description | Default
------ | ---- | ----------- | -------
docstring_style | "google" | "numpy" | "sphinx" | None | The docstring style to use. | "google"
docstring_options | dict[str, Any] | The options for the docstring parser. | {}
fallback_config: dict
¶collect(self, identifier, config)
¶Collect the documentation tree given an identifier and selection options.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
identifier |
str |
The dotted-path of a Python object available in the Python path. |
required |
config |
dict |
Selection options, used to alter the data collection done by |
required |
Exceptions:
| Type | Description |
|---|---|
CollectionError |
When there was a problem collecting the object documentation. |
Returns:
| Type | Description |
|---|---|
CollectorItem |
The collected object-tree. |
Source code in kiara/doc/mkdocstrings/collector.py
def collect(self, identifier: str, config: dict) -> CollectorItem: # noqa: WPS231
"""Collect the documentation tree given an identifier and selection options.
Arguments:
identifier: The dotted-path of a Python object available in the Python path.
config: Selection options, used to alter the data collection done by `pytkdocs`.
Raises:
CollectionError: When there was a problem collecting the object documentation.
Returns:
The collected object-tree.
"""
tokens = identifier.split(".")
if tokens[0] != "kiara_info":
return None
item_type = tokens[1]
item_id = ".".join(tokens[2:])
ctx: KiaraContextInfo = builtins.plugin_package_context_info # type: ignore
try:
item: KiaraInfoModel = ctx.get_info(item_type=item_type, item_id=item_id)
except Exception:
import traceback
traceback.print_exc()
raise CollectionError(f"Invalid id: {identifier}")
return {"obj": item, "identifier": identifier}
handler
¶
KiaraHandler (BaseHandler)
¶The kiara handler class.
Attributes:
| Name | Type | Description |
|---|---|---|
domain |
str |
The cross-documentation domain/language for this handler. |
enable_inventory |
bool |
Whether this handler is interested in enabling the creation
of the |
Source code in kiara/doc/mkdocstrings/handler.py
class KiaraHandler(BaseHandler):
"""The kiara handler class.
Attributes:
domain: The cross-documentation domain/language for this handler.
enable_inventory: Whether this handler is interested in enabling the creation
of the `objects.inv` Sphinx inventory file.
"""
domain: str = "kiara"
enable_inventory: bool = True
# load_inventory = staticmethod(inventory.list_object_urls)
#
# @classmethod
# def load_inventory(
# cls,
# in_file: typing.BinaryIO,
# url: str,
# base_url: typing.Optional[str] = None,
# **kwargs: typing.Any,
# ) -> typing.Iterator[typing.Tuple[str, str]]:
# """Yield items and their URLs from an inventory file streamed from `in_file`.
# This implements mkdocstrings' `load_inventory` "protocol" (see plugin.py).
# Arguments:
# in_file: The binary file-like object to read the inventory from.
# url: The URL that this file is being streamed from (used to guess `base_url`).
# base_url: The URL that this inventory's sub-paths are relative to.
# **kwargs: Ignore additional arguments passed from the config.
# Yields:
# Tuples of (item identifier, item URL).
# """
#
# print("XXXXXXXXXXXXXXXXXXXXXXXXXXXX")
#
# if base_url is None:
# base_url = posixpath.dirname(url)
#
# for item in Inventory.parse_sphinx(
# in_file, domain_filter=("py",)
# ).values(): # noqa: WPS526
# yield item.name, posixpath.join(base_url, item.uri)
get_handler(theme, custom_templates=None, **config)
¶Simply return an instance of PythonHandler.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
theme |
str |
The theme to use when rendering contents. |
required |
custom_templates |
Optional[str] |
Directory containing custom templates. |
None |
**config |
Any |
Configuration passed to the handler. |
{} |
Returns:
| Type | Description |
|---|---|
KiaraHandler |
An instance of |
Source code in kiara/doc/mkdocstrings/handler.py
def get_handler(
theme: str, # noqa: W0613 (unused argument config)
custom_templates: typing.Optional[str] = None,
**config: typing.Any,
) -> KiaraHandler:
"""Simply return an instance of `PythonHandler`.
Arguments:
theme: The theme to use when rendering contents.
custom_templates: Directory containing custom templates.
**config: Configuration passed to the handler.
Returns:
An instance of `PythonHandler`.
"""
if custom_templates is not None:
raise Exception("Custom templates are not supported for the kiara renderer.")
custom_templates = os.path.join(
KIARA_RESOURCES_FOLDER, "templates", "info_templates"
)
return KiaraHandler(
collector=KiaraCollector(),
renderer=KiaraInfoRenderer("kiara", theme, custom_templates),
)
renderer
¶
AliasResolutionError
¶Source code in kiara/doc/mkdocstrings/renderer.py
class AliasResolutionError:
pass
KiaraInfoRenderer (BaseRenderer)
¶Source code in kiara/doc/mkdocstrings/renderer.py
class KiaraInfoRenderer(BaseRenderer):
default_config: dict = {}
def get_anchors(
self, data: CollectorItem
) -> typing.List[str]: # noqa: D102 (ignore missing docstring)
if data is None:
return list()
return list([data["identifier"], data["kiara_id"], data["obj"].get_id()])
def render(self, data: typing.Dict[str, typing.Any], config: dict) -> str:
# final_config = ChainMap(config, self.default_config)
obj = data["obj"]
html = obj.create_html()
return html
default_config: dict
¶get_anchors(self, data)
¶Return the possible identifiers (HTML anchors) for a collected item.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
data |
Any |
The collected data. |
required |
Returns:
| Type | Description |
|---|---|
List[str] |
The HTML anchors (without '#'), or an empty tuple if this item doesn't have an anchor. |
Source code in kiara/doc/mkdocstrings/renderer.py
def get_anchors(
self, data: CollectorItem
) -> typing.List[str]: # noqa: D102 (ignore missing docstring)
if data is None:
return list()
return list([data["identifier"], data["kiara_id"], data["obj"].get_id()])
render(self, data, config)
¶Render a template using provided data and configuration options.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
data |
Dict[str, Any] |
The collected data to render. |
required |
config |
dict |
The rendering options. |
required |
Returns:
| Type | Description |
|---|---|
str |
The rendered template as HTML. |
Source code in kiara/doc/mkdocstrings/renderer.py
def render(self, data: typing.Dict[str, typing.Any], config: dict) -> str:
# final_config = ChainMap(config, self.default_config)
obj = data["obj"]
html = obj.create_html()
return html
exceptions
¶
FailedJobException (Exception)
¶
Source code in kiara/exceptions.py
class FailedJobException(Exception):
def __init__(self, job: "ActiveJob", msg: Optional[str] = None):
self.job: ActiveJob = job
if msg is None:
msg = "Job failed."
super().__init__(msg)
InvalidValuesException (Exception)
¶
Source code in kiara/exceptions.py
class InvalidValuesException(Exception):
def __init__(
self,
msg: Union[None, str, Exception] = None,
invalid_values: Mapping[str, str] = None,
):
if invalid_values is None:
invalid_values = {}
self.invalid_inputs: Mapping[str, str] = invalid_values
if msg is None:
if not self.invalid_inputs:
_msg = "Invalid values. No details available."
else:
msg_parts = []
for k, v in invalid_values.items():
msg_parts.append(f"{k}: {v}")
_msg = f"Invalid values: {', '.join(msg_parts)}"
elif isinstance(msg, Exception):
self._parent: Optional[Exception] = msg
_msg = str(msg)
else:
self._parent = None
_msg = msg
super().__init__(_msg)
def create_renderable(self, **config: Any) -> Table:
table = Table(box=box.SIMPLE, show_header=True)
table.add_column("field name", style="i")
table.add_column("[red]error[/red]")
for field_name, error in self.invalid_inputs.items():
row: List[RenderableType] = [field_name]
row.append(error)
table.add_row(*row)
return table
create_renderable(self, **config)
¶Source code in kiara/exceptions.py
def create_renderable(self, **config: Any) -> Table:
table = Table(box=box.SIMPLE, show_header=True)
table.add_column("field name", style="i")
table.add_column("[red]error[/red]")
for field_name, error in self.invalid_inputs.items():
row: List[RenderableType] = [field_name]
row.append(error)
table.add_row(*row)
return table
JobConfigException (Exception)
¶
Source code in kiara/exceptions.py
class JobConfigException(Exception):
def __init__(
self,
msg: Union[str, Exception],
manifest: "Manifest",
inputs: Mapping[str, Any],
):
self._manifest: Manifest = manifest
self._inputs: Mapping[str, Any] = inputs
if isinstance(msg, Exception):
self._parent: Optional[Exception] = msg
_msg = str(msg)
else:
self._parent = None
_msg = msg
super().__init__(_msg)
@property
def manifest(self) -> "Manifest":
return self._manifest
@property
def inputs(self) -> Mapping[str, Any]:
return self._inputs
KiaraException (Exception)
¶
Source code in kiara/exceptions.py
class KiaraException(Exception):
pass
KiaraModuleConfigException (Exception)
¶
Source code in kiara/exceptions.py
class KiaraModuleConfigException(Exception):
def __init__(
self,
msg: str,
module_cls: Type["KiaraModule"],
config: Mapping[str, Any],
parent: Optional[Exception] = None,
):
self._module_cls = module_cls
self._config = config
self._parent: Optional[Exception] = parent
if not msg.endswith("."):
_msg = msg + "."
else:
_msg = msg
super().__init__(_msg)
KiaraProcessingException (Exception)
¶
Source code in kiara/exceptions.py
class KiaraProcessingException(Exception):
def __init__(
self,
msg: Union[str, Exception],
module: Optional["KiaraModule"] = None,
inputs: Optional[Mapping[str, "Value"]] = None,
):
self._module: Optional["KiaraModule"] = module
self._inputs: Optional[Mapping[str, Value]] = inputs
if isinstance(msg, Exception):
self._parent: Optional[Exception] = msg
_msg = str(msg)
else:
self._parent = None
_msg = msg
super().__init__(_msg)
@property
def module(self) -> "KiaraModule":
return self._module # type: ignore
@property
def inputs(self) -> Mapping[str, "Value"]:
return self._inputs # type: ignore
@property
def parent_exception(self) -> Optional[Exception]:
return self._parent
KiaraValueException (Exception)
¶
Source code in kiara/exceptions.py
class KiaraValueException(Exception):
def __init__(
self,
data_type: Type["DataType"],
value_data: Any,
exception: Exception,
):
self._data_type: Type["DataType"] = data_type
self._value_data: Any = value_data
self._exception: Exception = exception
exc_msg = str(self._exception)
if not exc_msg:
exc_msg = "no details available"
super().__init__(f"Invalid value of type '{data_type._data_type_name}': {exc_msg}") # type: ignore
NoSuchExecutionTargetException (Exception)
¶
Source code in kiara/exceptions.py
class NoSuchExecutionTargetException(Exception):
def __init__(
self,
selected_target: str,
available_targets: Iterable[str],
msg: Optional[str] = None,
):
if msg is None:
msg = f"Specified run target '{selected_target}' is an operation, additional module configuration is not allowed."
self.avaliable_targets: Iterable[str] = available_targets
super().__init__(msg)
ValueTypeConfigException (Exception)
¶
Source code in kiara/exceptions.py
class ValueTypeConfigException(Exception):
def __init__(
self,
msg: str,
type_cls: Type["DataType"],
config: Mapping[str, Any],
parent: Optional[Exception] = None,
):
self._type_cls = type_cls
self._config = config
self._parent: Optional[Exception] = parent
if not msg.endswith("."):
_msg = msg + "."
else:
_msg = msg
super().__init__(_msg)
interfaces
special
¶
Implementation of interfaces for Kiara.
Functions¶
get_console()
¶
Get a global Console instance.
Returns:
| Type | Description |
|---|---|
Console |
A console instance. |
Source code in kiara/interfaces/__init__.py
def get_console() -> Console:
"""Get a global Console instance.
Returns:
Console: A console instance.
"""
global _console
if _console is None or True:
console_width = os.environ.get("CONSOLE_WIDTH", None)
width = None
if console_width:
try:
width = int(console_width)
except Exception:
pass
_console = Console(width=width)
return _console
Modules¶
cli
special
¶
A command-line interface for Kiara.
Modules¶
pipeline
special
¶commands
¶Pipeline-related subcommands for the cli.
get_pipeline_config(kiara_obj, pipeline_id_or_path)
¶Source code in kiara/interfaces/cli/pipeline/commands.py
def get_pipeline_config(kiara_obj: Kiara, pipeline_id_or_path: str) -> PipelineConfig:
if os.path.isfile(pipeline_id_or_path):
pc = PipelineConfig.from_file(pipeline_id_or_path, kiara=kiara_obj)
else:
operation: Operation = kiara_obj.operation_registry.get_operation(
pipeline_id_or_path
)
pipeline_module: PipelineModule = operation.module # type: ignore
if not pipeline_module.is_pipeline():
print()
print(
f"Specified operation id exists, but is not a pipeline: {pipeline_id_or_path}."
)
sys.exit(1)
pc = pipeline_module.config
return pc
run
¶The 'run' subcommand for the cli.
python_api
special
¶
logger
¶Classes¶
StoreValueResult (BaseModel)
pydantic-model
¶Source code in kiara/interfaces/python_api/__init__.py
class StoreValueResult(BaseModel):
value: Value = Field(description="The stored value.")
aliases: List[str] = Field(
description="The aliases that where assigned to the value when stored."
)
error: Optional[str] = Field(
description="An error that occured while trying to store."
)
StoreValuesResult (BaseModel)
pydantic-model
¶Source code in kiara/interfaces/python_api/__init__.py
class StoreValuesResult(BaseModel):
__root__: Dict[str, StoreValueResult]
def create_renderable(self, **config: Any) -> RenderableType:
table = Table(show_header=True, show_lines=False, box=box.SIMPLE)
table.add_column("field", style="b")
table.add_column("data type", style="i")
table.add_column("stored id", style="i")
table.add_column("alias(es)")
for field_name, value_result in self.__root__.items():
row = [
field_name,
str(value_result.value.value_schema.type),
str(value_result.value.value_id),
]
if value_result.aliases:
row.append(", ".join(value_result.aliases))
else:
row.append("")
table.add_row(*row)
return table
create_renderable(self, **config)
¶Source code in kiara/interfaces/python_api/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
table = Table(show_header=True, show_lines=False, box=box.SIMPLE)
table.add_column("field", style="b")
table.add_column("data type", style="i")
table.add_column("stored id", style="i")
table.add_column("alias(es)")
for field_name, value_result in self.__root__.items():
row = [
field_name,
str(value_result.value.value_schema.type),
str(value_result.value.value_id),
]
if value_result.aliases:
row.append(", ".join(value_result.aliases))
else:
row.append("")
table.add_row(*row)
return table
Modules¶
batch
¶
BatchOperation (BaseModel)
pydantic-model
¶Source code in kiara/interfaces/python_api/batch.py
class BatchOperation(BaseModel):
@classmethod
def from_file(
cls,
path: str,
kiara: Optional["Kiara"] = None,
):
data = get_data_from_file(path)
pipeline_id = data.get("pipeline_id", None)
if pipeline_id is None:
name = os.path.basename(path)
if name.endswith(".json"):
name = name[0:-5]
elif name.endswith(".yaml"):
name = name[0:-5]
data["pipeline_id"] = name
return cls.from_config(data=data, kiara=kiara)
@classmethod
def from_config(
cls,
data: Mapping[str, Any],
kiara: Optional["Kiara"],
):
data = dict(data)
inputs = data.pop("inputs", {})
save = data.pop("save", False)
pipeline_id = data.pop("pipeline_id", None)
if pipeline_id is None:
pipeline_id = str(uuid.uuid4())
if kiara is None:
kiara = Kiara.instance()
pipeline_config = PipelineConfig.from_config(
pipeline_id=pipeline_id, data=data, kiara=kiara
)
result = cls(pipeline_config=pipeline_config, inputs=inputs, save=save)
result._kiara = kiara
return result
alias: str = Field(description="The batch name/alias.")
pipeline_config: PipelineConfig = Field(
description="The configuration of the underlying pipeline."
)
inputs: Dict[str, Any] = Field(
description="The (base) inputs to use. Can be augmented before running the operation."
)
save: Dict[str, List[str]] = Field(
description="Configuration which values to save, under which alias(es).",
default_factory=dict,
)
_kiara: Kiara = PrivateAttr(default=None)
@root_validator(pre=True)
def add_alias(cls, values):
if not values.get("alias", None):
pc = values.get("pipeline_config", None)
if not pc:
raise ValueError("No pipeline config provided.")
if isinstance(pc, PipelineConfig):
alias = pc.pipeline_id
else:
alias = pc.get("pipeline_id", None)
values["alias"] = alias
return values
@validator("save", always=True, pre=True)
def validate_save(cls, save, values):
alias = values["alias"]
pipeline_config = values["pipeline_config"]
assert isinstance(pipeline_config, PipelineConfig)
if save in [False, None]:
save_new = {}
elif save is True:
field_names = pipeline_config.structure.pipeline_outputs_schema.keys()
save_new = create_save_config(field_names=field_names, aliases=alias)
elif isinstance(save, str):
field_names = pipeline_config.structure.pipeline_outputs_schema.keys()
save_new = create_save_config(field_names=field_names, aliases=save)
elif isinstance(save, Mapping):
save_new = save
field_names = pipeline_config.structure.pipeline_outputs_schema.keys()
save_new = create_save_config(field_names=field_names, aliases=save)
else:
raise ValueError(
f"Invalid type '{type(save)}' for 'save' attribute: must be None, bool, string or Mapping."
)
return save_new
def run(
self, save: bool = False, inputs: Optional[Mapping[str, Any]] = None
) -> ValueMap:
pipeline = Pipeline(
structure=self.pipeline_config.structure,
data_registry=self._kiara.data_registry,
)
pipeline_controller = SinglePipelineBatchController(
pipeline=pipeline, job_registry=self._kiara.job_registry
)
run_inputs = dict(self.inputs)
if inputs:
run_inputs.update(inputs)
pipeline.set_pipeline_inputs(inputs=run_inputs)
pipeline_controller.process_pipeline()
result = self._kiara.data_registry.load_values(
pipeline.get_current_pipeline_outputs()
)
if save:
self._kiara.save_values(values=result, alias_map=self.save)
return result
alias: str
pydantic-field
required
¶The batch name/alias.
inputs: Dict[str, Any]
pydantic-field
required
¶The (base) inputs to use. Can be augmented before running the operation.
pipeline_config: PipelineConfig
pydantic-field
required
¶The configuration of the underlying pipeline.
save: Dict[str, List[str]]
pydantic-field
¶Configuration which values to save, under which alias(es).
add_alias(values)
classmethod
¶Source code in kiara/interfaces/python_api/batch.py
@root_validator(pre=True)
def add_alias(cls, values):
if not values.get("alias", None):
pc = values.get("pipeline_config", None)
if not pc:
raise ValueError("No pipeline config provided.")
if isinstance(pc, PipelineConfig):
alias = pc.pipeline_id
else:
alias = pc.get("pipeline_id", None)
values["alias"] = alias
return values
from_config(data, kiara)
classmethod
¶Source code in kiara/interfaces/python_api/batch.py
@classmethod
def from_config(
cls,
data: Mapping[str, Any],
kiara: Optional["Kiara"],
):
data = dict(data)
inputs = data.pop("inputs", {})
save = data.pop("save", False)
pipeline_id = data.pop("pipeline_id", None)
if pipeline_id is None:
pipeline_id = str(uuid.uuid4())
if kiara is None:
kiara = Kiara.instance()
pipeline_config = PipelineConfig.from_config(
pipeline_id=pipeline_id, data=data, kiara=kiara
)
result = cls(pipeline_config=pipeline_config, inputs=inputs, save=save)
result._kiara = kiara
return result
from_file(path, kiara=None)
classmethod
¶Source code in kiara/interfaces/python_api/batch.py
@classmethod
def from_file(
cls,
path: str,
kiara: Optional["Kiara"] = None,
):
data = get_data_from_file(path)
pipeline_id = data.get("pipeline_id", None)
if pipeline_id is None:
name = os.path.basename(path)
if name.endswith(".json"):
name = name[0:-5]
elif name.endswith(".yaml"):
name = name[0:-5]
data["pipeline_id"] = name
return cls.from_config(data=data, kiara=kiara)
run(self, save=False, inputs=None)
¶Source code in kiara/interfaces/python_api/batch.py
def run(
self, save: bool = False, inputs: Optional[Mapping[str, Any]] = None
) -> ValueMap:
pipeline = Pipeline(
structure=self.pipeline_config.structure,
data_registry=self._kiara.data_registry,
)
pipeline_controller = SinglePipelineBatchController(
pipeline=pipeline, job_registry=self._kiara.job_registry
)
run_inputs = dict(self.inputs)
if inputs:
run_inputs.update(inputs)
pipeline.set_pipeline_inputs(inputs=run_inputs)
pipeline_controller.process_pipeline()
result = self._kiara.data_registry.load_values(
pipeline.get_current_pipeline_outputs()
)
if save:
self._kiara.save_values(values=result, alias_map=self.save)
return result
validate_save(save, values)
classmethod
¶Source code in kiara/interfaces/python_api/batch.py
@validator("save", always=True, pre=True)
def validate_save(cls, save, values):
alias = values["alias"]
pipeline_config = values["pipeline_config"]
assert isinstance(pipeline_config, PipelineConfig)
if save in [False, None]:
save_new = {}
elif save is True:
field_names = pipeline_config.structure.pipeline_outputs_schema.keys()
save_new = create_save_config(field_names=field_names, aliases=alias)
elif isinstance(save, str):
field_names = pipeline_config.structure.pipeline_outputs_schema.keys()
save_new = create_save_config(field_names=field_names, aliases=save)
elif isinstance(save, Mapping):
save_new = save
field_names = pipeline_config.structure.pipeline_outputs_schema.keys()
save_new = create_save_config(field_names=field_names, aliases=save)
else:
raise ValueError(
f"Invalid type '{type(save)}' for 'save' attribute: must be None, bool, string or Mapping."
)
return save_new
operation
¶
KiaraOperation
¶Source code in kiara/interfaces/python_api/operation.py
class KiaraOperation(object):
def __init__(
self,
kiara: "Kiara",
operation_name: str,
operation_config: Optional[Mapping[str, Any]] = None,
):
self._kiara: Kiara = kiara
self._operation_name: str = operation_name
if operation_config is None:
operation_config = {}
else:
operation_config = dict(operation_config)
self._operation_config: Dict[str, Any] = operation_config
self._inputs_raw: Dict[str, Any] = {}
self._operation: Optional[Operation] = None
self._inputs: Optional[ValueMap] = None
self._job_config: Optional[JobConfig] = None
self._queued_jobs: Dict[uuid.UUID, Dict[str, Any]] = {}
self._last_job: Optional[uuid.UUID] = None
self._results: Dict[uuid.UUID, ValueMap] = {}
def validate(self):
self.job_config # noqa
def _invalidate(self):
self._job_config = None
@property
def operation_inputs(self) -> ValueMap:
if self._inputs is not None:
return self._inputs
self._invalidate()
self._inputs = self._kiara.data_registry.create_valueset(
self._inputs_raw, self.operation.inputs_schema
)
return self._inputs
def set_input(self, field: Optional[str], value: Any = None):
if field is None:
if value is None:
self._inputs_raw.clear()
self._invalidate()
return
else:
if not isinstance(value, Mapping):
raise Exception(
"Can't set inputs dictionary (if no key is provided, value must be 'None' or of type 'Mapping')."
)
self._inputs_raw.clear()
self.set_inputs(**value)
self._invalidate()
return
else:
old = self._inputs_raw.get(field, None)
self._inputs_raw[field] = value
if old != value:
self._invalidate()
return
def set_inputs(self, **inputs: Any):
changed = False
for k, v in inputs.items():
old = self._inputs_raw.get(k, None)
self._inputs_raw[k] = v
if old != v:
changed = True
if changed:
self._invalidate()
return
@property
def operation_name(self) -> str:
return self._operation_name
@operation_name.setter
def operation_name(self, operation_name: str):
self._operation_name = operation_name
self._operation = None
@property
def operation_config(self) -> Mapping[str, Any]:
return self._operation_config
def set_operation_config_value(
self, key: Optional[str], value: Any = None
) -> Mapping[str, Any]:
if key is None:
if value is None:
old = bool(self._operation_config)
self._operation_config.clear()
if old:
self._operation = None
return self._operation_config
else:
try:
old_conf = self._operation_config
self._operation_config = dict(value)
if old_conf != self._operation_config:
self._operation = None
return self._operation_config
except Exception as e:
raise Exception(
f"Can't set configuration value dictionary (if no key is provided, value must be 'None' or of type 'Mapping'): {e}"
)
self._operation_config[key] = value
self._invalidate()
return self._operation_config
@property
def operation(self) -> "Operation":
if self._operation is not None:
return self._operation
self._invalidate()
module_or_operation = self._operation_name
operation: Optional[Operation] = None
if module_or_operation in self._kiara.operation_registry.operation_ids:
operation = self._kiara.operation_registry.get_operation(
module_or_operation
)
if self._operation_config:
raise Exception(
f"Specified run target '{module_or_operation}' is an operation, additional module configuration is not allowed."
)
elif module_or_operation in self._kiara.module_type_names:
manifest = Manifest(
module_type=module_or_operation, module_config=self._operation_config
)
module = self._kiara.create_module(manifest=manifest)
operation = Operation.create_from_module(module)
elif os.path.isfile(module_or_operation):
pipeline_config = PipelineConfig.from_file(
module_or_operation, kiara=self._kiara
)
manifest = self._kiara.create_manifest(
"pipeline", config=pipeline_config.dict()
)
module = self._kiara.create_module(manifest=manifest)
operation = Operation.create_from_module(module)
else:
raise Exception(
f"Can't assemble operation, invalid operation/module name: {module_or_operation}. Must be registered module or operation name, or file."
)
# manifest = Manifest(
# module_type=module_or_operation,
# module_config=self._operation_config,
# )
# module = self._kiara.create_module(manifest=manifest)
# operation = Operation.create_from_module(module=module)
if operation is None:
merged = set(self._kiara.module_type_names)
merged.update(self._kiara.operation_registry.operation_ids)
raise NoSuchExecutionTargetException(
selected_target=self.operation_name,
msg=f"Invalid run target name '{module_or_operation}'. Must be a path to a pipeline file, or one of the available modules/operations.",
available_targets=sorted(merged),
)
self._operation = operation
return self._operation
@property
def job_config(self) -> JobConfig:
if self._job_config is not None:
return self._job_config
self._job_config = self.operation.prepare_job_config(
kiara=self._kiara, inputs=self.operation_inputs
)
return self._job_config
def queue_job(self) -> uuid.UUID:
job_config = self.job_config
operation = self.operation
inputs = self.operation_inputs
job_id = self._kiara.job_registry.execute_job(job_config=job_config, wait=False)
self._queued_jobs[job_id] = {
"job_config": job_config,
"operation": operation,
"inputs": inputs,
}
self._last_job = job_id
return job_id
def retrieve_result(self, job_id: Optional[uuid.UUID] = None) -> ValueMap:
if job_id in self._results.keys():
assert job_id is not None
return self._results[job_id]
if job_id is None:
job_id = self._last_job
if job_id is None:
raise Exception("No job queued (yet).")
operation: Operation = self._queued_jobs[job_id]["operation"] # type: ignore
status = self._kiara.job_registry.get_job_status(job_id=job_id)
if status == JobStatus.FAILED:
job = self._kiara.job_registry.get_active_job(job_id=job_id)
raise FailedJobException(job=job)
outputs = self._kiara.job_registry.retrieve_result(job_id)
outputs = operation.process_job_outputs(outputs=outputs)
self._results[job_id] = outputs
return outputs
def save_result(
self,
job_id: Optional[uuid.UUID] = None,
aliases: Union[None, str, Mapping] = None,
) -> StoreValuesResult:
if job_id is None:
job_id = self._last_job
if job_id is None:
raise Exception("No job queued (yet).")
result = self.retrieve_result(job_id=job_id)
alias_map = create_save_config(field_names=result.field_names, aliases=aliases)
store_result = self._kiara.save_values(values=result, alias_map=alias_map)
self._kiara.job_registry.store_job_record(job_id=job_id)
return store_result
def create_renderable(self, **config: Any) -> RenderableType:
show_operation_name = config.get("show_operation_name", True)
show_operation_doc = config.get("show_operation_doc", True)
show_inputs_schema = config.get("show_inputs_schema", False)
items: List[Any] = []
if show_operation_name:
items.append(f"Operation: [bold]{self.operation_name}[/bold]")
if show_operation_doc and self.operation.doc.is_set:
items.append("")
items.append(Markdown(self.operation.doc.full_doc, style="i"))
if show_inputs_schema:
items.append("\nInputs:")
try:
op_inputs = self.operation_inputs
inputs: Any = create_value_map_status_renderable(op_inputs)
except InvalidValuesException as ive:
inputs = ive.create_renderable(**config)
except Exception as e:
inputs = f"[red bold]{e}[/red bold]"
items.append(inputs)
return Group(*items)
job_config: JobConfig
property
readonly
¶operation: Operation
property
readonly
¶operation_config: Mapping[str, Any]
property
readonly
¶operation_inputs: ValueMap
property
readonly
¶operation_name: str
property
writable
¶create_renderable(self, **config)
¶Source code in kiara/interfaces/python_api/operation.py
def create_renderable(self, **config: Any) -> RenderableType:
show_operation_name = config.get("show_operation_name", True)
show_operation_doc = config.get("show_operation_doc", True)
show_inputs_schema = config.get("show_inputs_schema", False)
items: List[Any] = []
if show_operation_name:
items.append(f"Operation: [bold]{self.operation_name}[/bold]")
if show_operation_doc and self.operation.doc.is_set:
items.append("")
items.append(Markdown(self.operation.doc.full_doc, style="i"))
if show_inputs_schema:
items.append("\nInputs:")
try:
op_inputs = self.operation_inputs
inputs: Any = create_value_map_status_renderable(op_inputs)
except InvalidValuesException as ive:
inputs = ive.create_renderable(**config)
except Exception as e:
inputs = f"[red bold]{e}[/red bold]"
items.append(inputs)
return Group(*items)
queue_job(self)
¶Source code in kiara/interfaces/python_api/operation.py
def queue_job(self) -> uuid.UUID:
job_config = self.job_config
operation = self.operation
inputs = self.operation_inputs
job_id = self._kiara.job_registry.execute_job(job_config=job_config, wait=False)
self._queued_jobs[job_id] = {
"job_config": job_config,
"operation": operation,
"inputs": inputs,
}
self._last_job = job_id
return job_id
retrieve_result(self, job_id=None)
¶Source code in kiara/interfaces/python_api/operation.py
def retrieve_result(self, job_id: Optional[uuid.UUID] = None) -> ValueMap:
if job_id in self._results.keys():
assert job_id is not None
return self._results[job_id]
if job_id is None:
job_id = self._last_job
if job_id is None:
raise Exception("No job queued (yet).")
operation: Operation = self._queued_jobs[job_id]["operation"] # type: ignore
status = self._kiara.job_registry.get_job_status(job_id=job_id)
if status == JobStatus.FAILED:
job = self._kiara.job_registry.get_active_job(job_id=job_id)
raise FailedJobException(job=job)
outputs = self._kiara.job_registry.retrieve_result(job_id)
outputs = operation.process_job_outputs(outputs=outputs)
self._results[job_id] = outputs
return outputs
save_result(self, job_id=None, aliases=None)
¶Source code in kiara/interfaces/python_api/operation.py
def save_result(
self,
job_id: Optional[uuid.UUID] = None,
aliases: Union[None, str, Mapping] = None,
) -> StoreValuesResult:
if job_id is None:
job_id = self._last_job
if job_id is None:
raise Exception("No job queued (yet).")
result = self.retrieve_result(job_id=job_id)
alias_map = create_save_config(field_names=result.field_names, aliases=aliases)
store_result = self._kiara.save_values(values=result, alias_map=alias_map)
self._kiara.job_registry.store_job_record(job_id=job_id)
return store_result
set_input(self, field, value=None)
¶Source code in kiara/interfaces/python_api/operation.py
def set_input(self, field: Optional[str], value: Any = None):
if field is None:
if value is None:
self._inputs_raw.clear()
self._invalidate()
return
else:
if not isinstance(value, Mapping):
raise Exception(
"Can't set inputs dictionary (if no key is provided, value must be 'None' or of type 'Mapping')."
)
self._inputs_raw.clear()
self.set_inputs(**value)
self._invalidate()
return
else:
old = self._inputs_raw.get(field, None)
self._inputs_raw[field] = value
if old != value:
self._invalidate()
return
set_inputs(self, **inputs)
¶Source code in kiara/interfaces/python_api/operation.py
def set_inputs(self, **inputs: Any):
changed = False
for k, v in inputs.items():
old = self._inputs_raw.get(k, None)
self._inputs_raw[k] = v
if old != v:
changed = True
if changed:
self._invalidate()
return
set_operation_config_value(self, key, value=None)
¶Source code in kiara/interfaces/python_api/operation.py
def set_operation_config_value(
self, key: Optional[str], value: Any = None
) -> Mapping[str, Any]:
if key is None:
if value is None:
old = bool(self._operation_config)
self._operation_config.clear()
if old:
self._operation = None
return self._operation_config
else:
try:
old_conf = self._operation_config
self._operation_config = dict(value)
if old_conf != self._operation_config:
self._operation = None
return self._operation_config
except Exception as e:
raise Exception(
f"Can't set configuration value dictionary (if no key is provided, value must be 'None' or of type 'Mapping'): {e}"
)
self._operation_config[key] = value
self._invalidate()
return self._operation_config
validate(self)
¶Source code in kiara/interfaces/python_api/operation.py
def validate(self):
self.job_config # noqa
utils
¶create_save_config(field_names, aliases)
¶Source code in kiara/interfaces/python_api/utils.py
def create_save_config(
field_names: Union[str, Iterable[str]],
aliases: Union[None, str, Iterable[str], Mapping[str, Any]],
) -> Dict[str, List[str]]:
if isinstance(field_names, str):
field_names = [field_names]
if aliases is None:
alias_map: Dict[str, List[str]] = {}
elif isinstance(aliases, str):
alias_map = {}
for field_name in field_names:
alias_map[field_name] = [f"{aliases}.{field_name}"]
elif isinstance(aliases, Mapping):
alias_map = {}
for field_name in aliases.keys():
if field_name in field_names:
if isinstance(aliases[field_name], str):
alias_map[field_name] = [aliases[field_name]]
else:
alias_map[field_name] = sorted(aliases[field_name])
else:
logger.warning(
"ignore.field_alias",
ignored_field_name=field_name,
reason="field name not in results",
available_field_names=sorted(field_names),
)
continue
else:
raise Exception(
f"Invalid type '{type(aliases)}' for aliases parameter, must be string or mapping."
)
return alias_map
models
special
¶
Classes¶
KiaraModel (ABC, BaseModel, JupyterMixin)
pydantic-model
¶
Base class that all models in kiara inherit from.
This class provides utility functions for things like rendering the model on terminal or as html, integration into a tree hierarchy of the overall kiara context, hashing, etc.
Source code in kiara/models/__init__.py
class KiaraModel(ABC, BaseModel, JupyterMixin):
"""Base class that all models in kiara inherit from.
This class provides utility functions for things like rendering the model on terminal or as html, integration into
a tree hierarchy of the overall kiara context, hashing, etc.
"""
__slots__ = ["__weakref__"]
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
# allow_mutation = False
@classmethod
def get_schema_hash(cls) -> int:
if cls._schema_hash_cache is not None:
return cls._schema_hash_cache
obj = cls.schema_json()
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
cls._schema_hash_cache = h[obj]
return cls._schema_hash_cache
_graph_cache: Optional[nx.DiGraph] = PrivateAttr(default=None)
_subcomponent_names_cache: Optional[List[str]] = PrivateAttr(default=None)
_dynamic_subcomponents: Dict[str, "KiaraModel"] = PrivateAttr(default_factory=dict)
_id_cache: Optional[str] = PrivateAttr(default=None)
_category_id_cache: Optional[str] = PrivateAttr(default=None)
_schema_hash_cache: ClassVar = None
_hash_cache: Optional[int] = PrivateAttr(default=None)
_size_cache: Optional[int] = PrivateAttr(default=None)
@abstractmethod
def _retrieve_id(self) -> str:
"""Retrieve the unique id (with its category) of this model."""
@abstractmethod
def _retrieve_category_id(self) -> str:
"""Return the id of the category this model is part of."""
@abstractmethod
def _retrieve_data_to_hash(self) -> Any:
"""Return data important for hashing this model instance. Implemented by sub-classes.
This returns the relevant data that makes this model unique, excluding any secondary metadata that is not
necessary for this model to be used functionally. Like for example documentation.
"""
@property
def model_id(self) -> str:
"""The unique id of this model, within its category."""
if self._id_cache is None:
self._id_cache = self._retrieve_id()
return self._id_cache
@property
def category_id(self) -> str:
"""The id of the category of this model."""
if self._category_id_cache is None:
self._category_id_cache = self._retrieve_category_id()
return self._category_id_cache
@property
def model_data_hash(self) -> int:
"""A hash for this model."""
if self._hash_cache is not None:
return self._hash_cache
obj = self._retrieve_data_to_hash()
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
self._hash_cache = h[obj]
return self._hash_cache
@property
def model_size(self) -> int:
if self._size_cache is not None:
return self._size_cache
self._size_cache = sys.getsizeof(self.dict())
return self._size_cache
# ==========================================================================================
# subcomponent related methods
@property
def subcomponent_keys(self) -> Iterable[str]:
"""The keys of available sub-components of this model."""
if self._subcomponent_names_cache is None:
self._subcomponent_names_cache = sorted(self._retrieve_subcomponent_keys())
return self._subcomponent_names_cache
@property
def subcomponent_tree(self) -> Optional[nx.DiGraph]:
"""A tree structure, containing all sub-components (and their subcomponents) of this model."""
if not self.subcomponent_keys:
return None
if self._graph_cache is None:
self._graph_cache = assemble_subcomponent_tree(self)
return self._graph_cache
def get_subcomponent(self, path: str) -> "KiaraModel":
"""Retrieve the subcomponent identified by the specified path."""
if path not in self._dynamic_subcomponents.keys():
self._dynamic_subcomponents[path] = self._retrieve_subcomponent(path=path)
return self._dynamic_subcomponents[path]
def find_subcomponents(self, category: str) -> Dict[str, "KiaraModel"]:
"""Find and return all subcomponents of this model that are member of the specified category."""
tree = self.subcomponent_tree
if tree is None:
raise Exception(f"No subcomponents found for category: {category}")
result = {}
for node_id, node in tree.nodes.items():
if not hasattr(node["obj"], "get_category_alias"):
raise NotImplementedError()
if category != node["obj"].get_category_alias():
continue
n_id = node_id[9:] # remove the __self__. token
result[n_id] = node["obj"]
return result
def _retrieve_subcomponent_keys(self) -> Iterable[str]:
"""Retrieve the keys of all subcomponents of this model.
Can be overwritten in sub-classes, by default it tries to automatically determine the subcomponents.
"""
return retrieve_data_subcomponent_keys(self)
def _retrieve_subcomponent(self, path: str) -> "KiaraModel":
"""Retrieve the subcomponent under the specified path.
Can be overwritten in sub-classes, by default it tries to automatically determine the subcomponents.
"""
m = get_subcomponent_from_model(self, path=path)
return m
# ==========================================================================================
# model rendering related methods
def create_panel(self, title: str = None, **config: Any) -> Panel:
rend = self.create_renderable(**config)
return Panel(rend, box=box.ROUNDED, title=title, title_align="left")
def create_renderable(self, **config: Any) -> RenderableType:
from kiara.utils.output import extract_renderable
include = config.get("include", None)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("Key", style="i")
table.add_column("Value")
for k in self.__fields__.keys():
if include is not None and k not in include:
continue
attr = getattr(self, k)
v = extract_renderable(attr)
table.add_row(k, v)
return table
def create_html(self, **config) -> str:
r = self.create_renderable()
if hasattr(r, "_repr_mimebundle_"):
mime_bundle = r._repr_mimebundle_(include=[], exclude=[]) # type: ignore
else:
raise NotImplementedError(
f"Type '{self.__class__}' can't be rendered as html (yet)."
)
return mime_bundle["text/html"]
def as_dict_with_schema(self) -> Dict[str, Dict[str, Any]]:
return {"data": self.dict(), "schema": self.schema()}
def as_json_with_schema(self) -> str:
data_json = self.json()
schema_json = self.schema_json()
return '{"data": ' + data_json + ', "schema": ' + schema_json + "}"
def __hash__(self):
return self.model_data_hash
def __eq__(self, other):
if self.__class__ != other.__class__:
return False
else:
return (self.model_id, self.model_data_hash) == (
other.model_id,
other.model_data_hash,
)
def __repr__(self):
return f"{self.__class__.__name__}(model_id={self.model_id}, category={self.category_id}, fields=[{', '.join(self.__fields__.keys())}])"
def __str__(self):
return self.__repr__()
def __rich_console__(
self, console: Console, options: ConsoleOptions
) -> RenderResult:
yield self.create_renderable()
Attributes¶
category_id: str
property
readonly
¶The id of the category of this model.
model_data_hash: int
property
readonly
¶A hash for this model.
model_id: str
property
readonly
¶The unique id of this model, within its category.
model_size: int
property
readonly
¶subcomponent_keys: Iterable[str]
property
readonly
¶The keys of available sub-components of this model.
subcomponent_tree: Optional[networkx.classes.digraph.DiGraph]
property
readonly
¶A tree structure, containing all sub-components (and their subcomponents) of this model.
Config
¶Source code in kiara/models/__init__.py
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
# allow_mutation = False
json_loads
¶json_dumps(v, *, default=None, **args)
¶Source code in kiara/models/__init__.py
def orjson_dumps(v, *, default=None, **args):
# orjson.dumps returns bytes, to match standard json.dumps we need to decode
try:
return orjson.dumps(v, default=default, **args).decode()
except Exception as e:
if is_debug():
print(f"Error dumping json data: {e}")
from kiara import dbg
dbg(v)
raise e
Methods¶
as_dict_with_schema(self)
¶Source code in kiara/models/__init__.py
def as_dict_with_schema(self) -> Dict[str, Dict[str, Any]]:
return {"data": self.dict(), "schema": self.schema()}
as_json_with_schema(self)
¶Source code in kiara/models/__init__.py
def as_json_with_schema(self) -> str:
data_json = self.json()
schema_json = self.schema_json()
return '{"data": ' + data_json + ', "schema": ' + schema_json + "}"
create_html(self, **config)
¶Source code in kiara/models/__init__.py
def create_html(self, **config) -> str:
r = self.create_renderable()
if hasattr(r, "_repr_mimebundle_"):
mime_bundle = r._repr_mimebundle_(include=[], exclude=[]) # type: ignore
else:
raise NotImplementedError(
f"Type '{self.__class__}' can't be rendered as html (yet)."
)
return mime_bundle["text/html"]
create_panel(self, title=None, **config)
¶Source code in kiara/models/__init__.py
def create_panel(self, title: str = None, **config: Any) -> Panel:
rend = self.create_renderable(**config)
return Panel(rend, box=box.ROUNDED, title=title, title_align="left")
create_renderable(self, **config)
¶Source code in kiara/models/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
from kiara.utils.output import extract_renderable
include = config.get("include", None)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("Key", style="i")
table.add_column("Value")
for k in self.__fields__.keys():
if include is not None and k not in include:
continue
attr = getattr(self, k)
v = extract_renderable(attr)
table.add_row(k, v)
return table
find_subcomponents(self, category)
¶Find and return all subcomponents of this model that are member of the specified category.
Source code in kiara/models/__init__.py
def find_subcomponents(self, category: str) -> Dict[str, "KiaraModel"]:
"""Find and return all subcomponents of this model that are member of the specified category."""
tree = self.subcomponent_tree
if tree is None:
raise Exception(f"No subcomponents found for category: {category}")
result = {}
for node_id, node in tree.nodes.items():
if not hasattr(node["obj"], "get_category_alias"):
raise NotImplementedError()
if category != node["obj"].get_category_alias():
continue
n_id = node_id[9:] # remove the __self__. token
result[n_id] = node["obj"]
return result
get_schema_hash()
classmethod
¶Source code in kiara/models/__init__.py
@classmethod
def get_schema_hash(cls) -> int:
if cls._schema_hash_cache is not None:
return cls._schema_hash_cache
obj = cls.schema_json()
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
cls._schema_hash_cache = h[obj]
return cls._schema_hash_cache
get_subcomponent(self, path)
¶Retrieve the subcomponent identified by the specified path.
Source code in kiara/models/__init__.py
def get_subcomponent(self, path: str) -> "KiaraModel":
"""Retrieve the subcomponent identified by the specified path."""
if path not in self._dynamic_subcomponents.keys():
self._dynamic_subcomponents[path] = self._retrieve_subcomponent(path=path)
return self._dynamic_subcomponents[path]
Modules¶
aliases
special
¶
VALUE_ALIAS_SEPARATOR
¶logger
¶Classes¶
AliasValueMap (ValueMap)
pydantic-model
¶Source code in kiara/models/aliases/__init__.py
class AliasValueMap(ValueMap):
alias: Optional[str] = Field(description="This maps own (full) alias.")
version: int = Field(description="The version of this map (in this maps parent).")
created: Optional[datetime.datetime] = Field(
description="The time this map was created."
)
assoc_schema: Optional[ValueSchema] = Field(
description="The schema for this maps associated value."
)
assoc_value: Optional[uuid.UUID] = Field(
description="The value that is associated with this map."
)
value_items: Dict[str, Dict[int, "AliasValueMap"]] = Field(
description="The values contained in this set.", default_factory=dict
)
_data_registry: "DataRegistry" = PrivateAttr(default=None)
_schema_locked: bool = PrivateAttr(default=False)
_auto_schema: bool = PrivateAttr(default=True)
_is_stored: bool = PrivateAttr(default=False)
def _retrieve_id(self) -> str:
return str(uuid.uuid4())
def _retrieve_category_id(self) -> str:
return VALUES_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
raise NotImplementedError()
@property
def is_stored(self) -> bool:
return self._is_stored
def get_child_map(
self, field_name: str, version: Optional[str] = None
) -> Optional["AliasValueMap"]:
"""Get the child map for the specified field / version combination.
Raises an error if the child field does not exist. Returns 'None' if not value is set yet (but schema is).
"""
if version is not None:
raise NotImplementedError()
if VALUE_ALIAS_SEPARATOR not in field_name:
if self.values_schema.get(field_name, None) is None:
raise KeyError(
f"No field name '{field_name}'. Available fields: {', '.join(self.values_schema.keys())}"
)
field_items = self.value_items[field_name]
if not field_items:
return None
max_version = max(field_items.keys())
item = field_items[max_version]
return item
else:
child, rest = field_name.split(VALUE_ALIAS_SEPARATOR, maxsplit=1)
if child not in self.values_schema.keys():
raise Exception(
f"No field name '{child}'. Available fields: {', '.join(self.values_schema.keys())}"
)
child_map = self.get_child_map(child)
assert child_map is not None
return child_map.get_child_map(rest)
def get_value_obj(self, field_name: str) -> Value:
item = self.get_child_map(field_name=field_name)
if item is None:
return self._data_registry.NONE_VALUE
if item.assoc_value is None:
raise Exception(f"No value associated for field '{field_name}'.")
return self._data_registry.get_value(value_id=item.assoc_value)
def get_value_id(self, field_name: str) -> uuid.UUID:
item = self.get_child_map(field_name=field_name)
if item is None:
return NONE_VALUE_ID
else:
return item.assoc_value if item.assoc_value is not None else NONE_VALUE_ID
def get_all_value_ids(
self,
) -> Dict[str, uuid.UUID]:
result: Dict[str, uuid.UUID] = {}
for k in self.values_schema.keys():
v_id = self.get_value_id(field_name=k)
if v_id is None:
v_id = NONE_VALUE_ID
result[k] = v_id
return result
def set_value(self, field_name: str, data: Any) -> None:
assert VALUE_ALIAS_SEPARATOR not in field_name
value = self._data_registry.register_data(data)
self.set_alias(alias=field_name, value_id=value.value_id)
def set_alias_schema(self, alias: str, schema: ValueSchema):
if self._schema_locked:
raise Exception(f"Can't add schema for alias '{alias}': schema locked.")
if VALUE_ALIAS_SEPARATOR not in alias:
self._set_local_field_schema(field_name=alias, schema=schema)
else:
child, rest = alias.split(VALUE_ALIAS_SEPARATOR, maxsplit=1)
if child in self.values_schema.keys():
child_map = self.get_child_map(child)
else:
self._set_local_field_schema(
field_name=child, schema=ValueSchema(type="none")
)
child_map = self.set_alias(alias=child, value_id=None)
assert child_map is not None
child_map.set_alias_schema(alias=rest, schema=schema)
def _set_local_field_schema(self, field_name: str, schema: ValueSchema):
assert field_name is not None
if VALUE_ALIAS_SEPARATOR in field_name:
raise Exception(
f"Can't add schema, field name has alias separator in name: {field_name}. This is most likely a bug."
)
if field_name in self.values_schema.keys():
raise Exception(
f"Can't set alias schema for '{field_name}' to map: schema already set."
)
try:
items = self.get_child_map(field_name)
if items is not None:
raise Exception(
f"Can't set schema for field '{field_name}': already at least one child set for this field."
)
except KeyError:
pass
self.values_schema[field_name] = schema
self.value_items[field_name] = {}
def get_alias(self, alias: str) -> Optional["AliasValueMap"]:
if VALUE_ALIAS_SEPARATOR not in alias:
if "@" in alias:
raise NotImplementedError()
child_map = self.get_child_map(alias)
if child_map is None:
return None
return child_map
else:
child, rest = alias.split(VALUE_ALIAS_SEPARATOR, maxsplit=1)
if "@" in child:
raise NotImplementedError()
child_map = self.get_child_map(field_name=child)
if child_map is None:
return None
return child_map.get_alias(rest)
def set_aliases(self, **aliases) -> Mapping[str, "AliasValueMap"]:
result = {}
for k, v in aliases.items():
r = self.set_alias(alias=k, value_id=v)
result[k] = r
return result
def set_alias(self, alias: str, value_id: Optional[uuid.UUID]) -> "AliasValueMap":
if VALUE_ALIAS_SEPARATOR not in alias:
child = None
field_name: Optional[str] = alias
rest = None
else:
child, rest = alias.split(VALUE_ALIAS_SEPARATOR, maxsplit=1)
field_name = None
if child is None:
# means we are setting the alias in this map
assert field_name is not None
new_map = self._set_local_value_item(
field_name=field_name, value_id=value_id
)
return new_map
else:
# means we are dealing with an intermediate alias map
assert rest is not None
assert child is not None
assert field_name is None
if child not in self.value_items.keys():
if not self._auto_schema:
raise Exception(
f"Can't set alias '{alias}', no schema set for field: '{child}'."
)
else:
self.set_alias_schema(alias=child, schema=ValueSchema(type="any"))
field_item: Optional[AliasValueMap] = None
try:
field_item = self.get_child_map(field_name=child)
except KeyError:
pass
if self.alias:
new_alias = f"{self.alias}.{child}"
else:
new_alias = child
if field_item is None:
new_version = 0
schemas = {}
self.value_items[child] = {}
else:
max_version = len(field_item.keys())
new_version = max_version + 1
assert field_item.alias == new_alias
assert field_item.version == max_version
schemas = field_item.values_schema
new_map = AliasValueMap(
alias=new_alias,
version=new_version,
assoc_schema=self.values_schema[child],
assoc_value=None,
values_schema=schemas,
)
new_map._data_registry = self._data_registry
self.value_items[child][new_version] = new_map
new_map.set_alias(alias=rest, value_id=value_id)
return new_map
def _set_local_value_item(
self, field_name: str, value_id: Optional[uuid.UUID] = None
) -> "AliasValueMap":
assert VALUE_ALIAS_SEPARATOR not in field_name
value: Optional[Value] = None
if value_id is not None:
value = self._data_registry.get_value(value_id=value_id)
assert value is not None
assert value.value_id == value_id
if field_name not in self.values_schema.keys():
if not self._auto_schema:
raise Exception(
f"Can't add value for field '{field_name}': field not in schema."
)
else:
if value_id is None:
value_schema = ValueSchema(type="none")
else:
value_schema = value.value_schema # type: ignore
self.set_alias_schema(alias=field_name, schema=value_schema)
field_items = self.value_items.get(field_name, None)
if not field_items:
assert field_items is not None
new_version = 0
values_schema = {}
else:
max_version = max(field_items.keys())
current_map = field_items[max_version]
if value_id == current_map.assoc_value:
logger.debug(
"set_field.skip",
value_id=None,
reason=f"Same value id: {value_id}",
)
return current_map
# TODO: check schema
new_version = max(field_items.keys()) + 1
values_schema = current_map.values_schema
if self.alias:
new_alias = f"{self.alias}.{field_name}"
else:
new_alias = field_name
new_map = AliasValueMap(
alias=new_alias,
version=new_version,
assoc_schema=self.values_schema[field_name],
assoc_value=value_id,
values_schema=values_schema,
)
new_map._data_registry = self._data_registry
self.value_items[field_name][new_version] = new_map
return new_map
def print_tree(self):
t = self.get_tree("base")
terminal_print(t)
def get_tree(self, base_name: str) -> Tree:
if self.assoc_schema:
type_name = self.assoc_schema.type
else:
type_name = "none"
if type_name == "none":
type_str = ""
else:
type_str = f" ({type_name})"
tree = Tree(f"{base_name}{type_str}")
if self.assoc_value:
data = tree.add("__data__")
value = self._data_registry.get_value(self.assoc_value)
data.add(str(value.data))
for field_name, schema in self.values_schema.items():
alias = self.get_alias(alias=field_name)
if alias is not None:
tree.add(alias.get_tree(base_name=field_name))
else:
if schema.type == "none":
type_str = ""
else:
type_str = f" ({schema.type})"
tree.add(f"{field_name}{type_str}")
return tree
def __repr__(self):
return f"AliasMap(assoc_value={self.assoc_value}, field_names={self.value_items.keys()})"
def __str__(self):
return self.__repr__()
alias: str
pydantic-field
¶This maps own (full) alias.
assoc_schema: ValueSchema
pydantic-field
¶The schema for this maps associated value.
assoc_value: UUID
pydantic-field
¶The value that is associated with this map.
created: datetime
pydantic-field
¶The time this map was created.
is_stored: bool
property
readonly
¶value_items: Dict[str, Dict[int, AliasValueMap]]
pydantic-field
¶The values contained in this set.
version: int
pydantic-field
required
¶The version of this map (in this maps parent).
get_alias(self, alias)
¶Source code in kiara/models/aliases/__init__.py
def get_alias(self, alias: str) -> Optional["AliasValueMap"]:
if VALUE_ALIAS_SEPARATOR not in alias:
if "@" in alias:
raise NotImplementedError()
child_map = self.get_child_map(alias)
if child_map is None:
return None
return child_map
else:
child, rest = alias.split(VALUE_ALIAS_SEPARATOR, maxsplit=1)
if "@" in child:
raise NotImplementedError()
child_map = self.get_child_map(field_name=child)
if child_map is None:
return None
return child_map.get_alias(rest)
get_all_value_ids(self)
¶Source code in kiara/models/aliases/__init__.py
def get_all_value_ids(
self,
) -> Dict[str, uuid.UUID]:
result: Dict[str, uuid.UUID] = {}
for k in self.values_schema.keys():
v_id = self.get_value_id(field_name=k)
if v_id is None:
v_id = NONE_VALUE_ID
result[k] = v_id
return result
get_child_map(self, field_name, version=None)
¶Get the child map for the specified field / version combination.
Raises an error if the child field does not exist. Returns 'None' if not value is set yet (but schema is).
Source code in kiara/models/aliases/__init__.py
def get_child_map(
self, field_name: str, version: Optional[str] = None
) -> Optional["AliasValueMap"]:
"""Get the child map for the specified field / version combination.
Raises an error if the child field does not exist. Returns 'None' if not value is set yet (but schema is).
"""
if version is not None:
raise NotImplementedError()
if VALUE_ALIAS_SEPARATOR not in field_name:
if self.values_schema.get(field_name, None) is None:
raise KeyError(
f"No field name '{field_name}'. Available fields: {', '.join(self.values_schema.keys())}"
)
field_items = self.value_items[field_name]
if not field_items:
return None
max_version = max(field_items.keys())
item = field_items[max_version]
return item
else:
child, rest = field_name.split(VALUE_ALIAS_SEPARATOR, maxsplit=1)
if child not in self.values_schema.keys():
raise Exception(
f"No field name '{child}'. Available fields: {', '.join(self.values_schema.keys())}"
)
child_map = self.get_child_map(child)
assert child_map is not None
return child_map.get_child_map(rest)
get_tree(self, base_name)
¶Source code in kiara/models/aliases/__init__.py
def get_tree(self, base_name: str) -> Tree:
if self.assoc_schema:
type_name = self.assoc_schema.type
else:
type_name = "none"
if type_name == "none":
type_str = ""
else:
type_str = f" ({type_name})"
tree = Tree(f"{base_name}{type_str}")
if self.assoc_value:
data = tree.add("__data__")
value = self._data_registry.get_value(self.assoc_value)
data.add(str(value.data))
for field_name, schema in self.values_schema.items():
alias = self.get_alias(alias=field_name)
if alias is not None:
tree.add(alias.get_tree(base_name=field_name))
else:
if schema.type == "none":
type_str = ""
else:
type_str = f" ({schema.type})"
tree.add(f"{field_name}{type_str}")
return tree
get_value_id(self, field_name)
¶Source code in kiara/models/aliases/__init__.py
def get_value_id(self, field_name: str) -> uuid.UUID:
item = self.get_child_map(field_name=field_name)
if item is None:
return NONE_VALUE_ID
else:
return item.assoc_value if item.assoc_value is not None else NONE_VALUE_ID
get_value_obj(self, field_name)
¶Source code in kiara/models/aliases/__init__.py
def get_value_obj(self, field_name: str) -> Value:
item = self.get_child_map(field_name=field_name)
if item is None:
return self._data_registry.NONE_VALUE
if item.assoc_value is None:
raise Exception(f"No value associated for field '{field_name}'.")
return self._data_registry.get_value(value_id=item.assoc_value)
print_tree(self)
¶Source code in kiara/models/aliases/__init__.py
def print_tree(self):
t = self.get_tree("base")
terminal_print(t)
set_alias(self, alias, value_id)
¶Source code in kiara/models/aliases/__init__.py
def set_alias(self, alias: str, value_id: Optional[uuid.UUID]) -> "AliasValueMap":
if VALUE_ALIAS_SEPARATOR not in alias:
child = None
field_name: Optional[str] = alias
rest = None
else:
child, rest = alias.split(VALUE_ALIAS_SEPARATOR, maxsplit=1)
field_name = None
if child is None:
# means we are setting the alias in this map
assert field_name is not None
new_map = self._set_local_value_item(
field_name=field_name, value_id=value_id
)
return new_map
else:
# means we are dealing with an intermediate alias map
assert rest is not None
assert child is not None
assert field_name is None
if child not in self.value_items.keys():
if not self._auto_schema:
raise Exception(
f"Can't set alias '{alias}', no schema set for field: '{child}'."
)
else:
self.set_alias_schema(alias=child, schema=ValueSchema(type="any"))
field_item: Optional[AliasValueMap] = None
try:
field_item = self.get_child_map(field_name=child)
except KeyError:
pass
if self.alias:
new_alias = f"{self.alias}.{child}"
else:
new_alias = child
if field_item is None:
new_version = 0
schemas = {}
self.value_items[child] = {}
else:
max_version = len(field_item.keys())
new_version = max_version + 1
assert field_item.alias == new_alias
assert field_item.version == max_version
schemas = field_item.values_schema
new_map = AliasValueMap(
alias=new_alias,
version=new_version,
assoc_schema=self.values_schema[child],
assoc_value=None,
values_schema=schemas,
)
new_map._data_registry = self._data_registry
self.value_items[child][new_version] = new_map
new_map.set_alias(alias=rest, value_id=value_id)
return new_map
set_alias_schema(self, alias, schema)
¶Source code in kiara/models/aliases/__init__.py
def set_alias_schema(self, alias: str, schema: ValueSchema):
if self._schema_locked:
raise Exception(f"Can't add schema for alias '{alias}': schema locked.")
if VALUE_ALIAS_SEPARATOR not in alias:
self._set_local_field_schema(field_name=alias, schema=schema)
else:
child, rest = alias.split(VALUE_ALIAS_SEPARATOR, maxsplit=1)
if child in self.values_schema.keys():
child_map = self.get_child_map(child)
else:
self._set_local_field_schema(
field_name=child, schema=ValueSchema(type="none")
)
child_map = self.set_alias(alias=child, value_id=None)
assert child_map is not None
child_map.set_alias_schema(alias=rest, schema=schema)
set_aliases(self, **aliases)
¶Source code in kiara/models/aliases/__init__.py
def set_aliases(self, **aliases) -> Mapping[str, "AliasValueMap"]:
result = {}
for k, v in aliases.items():
r = self.set_alias(alias=k, value_id=v)
result[k] = r
return result
set_value(self, field_name, data)
¶Source code in kiara/models/aliases/__init__.py
def set_value(self, field_name: str, data: Any) -> None:
assert VALUE_ALIAS_SEPARATOR not in field_name
value = self._data_registry.register_data(data)
self.set_alias(alias=field_name, value_id=value.value_id)
documentation
¶
Classes¶
AuthorModel (BaseModel)
pydantic-model
¶Source code in kiara/models/documentation.py
class AuthorModel(BaseModel):
name: str = Field(description="The full name of the author.")
email: Optional[EmailStr] = Field(
description="The email address of the author", default=None
)
AuthorsMetadataModel (KiaraModel)
pydantic-model
¶Source code in kiara/models/documentation.py
class AuthorsMetadataModel(KiaraModel):
_metadata_key = "origin"
@classmethod
def from_class(cls, item_cls: Type):
data = get_metadata_for_python_module_or_class(item_cls) # type: ignore
merged = merge_dicts(*data)
return cls.parse_obj(merged)
authors: List[AuthorModel] = Field(
description="The authors/creators of this item.", default_factory=list
)
def _retrieve_id(self) -> str:
return str(uuid.uuid4())
def _retrieve_category_id(self) -> str:
return AUTHORS_METADATA_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.authors
def create_renderable(self, **config: Any) -> RenderableType:
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("Name")
table.add_column("Email", style="i")
for author in reversed(self.authors):
if author.email:
authors: Tuple[str, Union[str, EmailStr]] = (author.name, author.email)
else:
authors = (author.name, "")
table.add_row(*authors)
return table
authors: List[kiara.models.documentation.AuthorModel]
pydantic-field
¶The authors/creators of this item.
create_renderable(self, **config)
¶Source code in kiara/models/documentation.py
def create_renderable(self, **config: Any) -> RenderableType:
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("Name")
table.add_column("Email", style="i")
for author in reversed(self.authors):
if author.email:
authors: Tuple[str, Union[str, EmailStr]] = (author.name, author.email)
else:
authors = (author.name, "")
table.add_row(*authors)
return table
from_class(item_cls)
classmethod
¶Source code in kiara/models/documentation.py
@classmethod
def from_class(cls, item_cls: Type):
data = get_metadata_for_python_module_or_class(item_cls) # type: ignore
merged = merge_dicts(*data)
return cls.parse_obj(merged)
ContextMetadataModel (KiaraModel)
pydantic-model
¶Source code in kiara/models/documentation.py
class ContextMetadataModel(KiaraModel):
@classmethod
def from_class(cls, item_cls: Type):
data = get_metadata_for_python_module_or_class(item_cls) # type: ignore
merged = merge_dicts(*data)
return cls.parse_obj(merged)
_metadata_key = "properties"
references: Dict[str, LinkModel] = Field(
description="References for the item.", default_factory=dict
)
tags: List[str] = Field(
description="A list of tags for the item.", default_factory=list
)
labels: Dict[str, str] = Field(
description="A list of labels for the item.", default_factory=list
)
def _retrieve_id(self) -> str:
return str(uuid.uuid4())
def _retrieve_category_id(self) -> str:
return CONTEXT_METADATA_CATEOGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
def create_renderable(self, **config: Any) -> RenderableType:
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("Key", style="i")
table.add_column("Value")
if self.tags:
table.add_row("Tags", ", ".join(self.tags))
if self.labels:
labels = []
for k, v in self.labels.items():
labels.append(f"[i]{k}[/i]: {v}")
table.add_row("Labels", "\n".join(labels))
if self.references:
references = []
for _k, _v in self.references.items():
link = f"[link={_v.url}]{_v.url}[/link]"
references.append(f"[i]{_k}[/i]: {link}")
table.add_row("References", "\n".join(references))
return table
def add_reference(
self,
ref_type: str,
url: str,
desc: Optional[str] = None,
force: bool = False,
):
if ref_type in self.references.keys() and not force:
raise Exception(f"Reference of type '{ref_type}' already present.")
link = LinkModel(url=url, desc=desc)
self.references[ref_type] = link
def get_url_for_reference(self, ref: str) -> Optional[str]:
link = self.references.get(ref, None)
if not link:
return None
return link.url
labels: Dict[str, str]
pydantic-field
¶A list of labels for the item.
references: Dict[str, kiara.models.documentation.LinkModel]
pydantic-field
¶References for the item.
tags: List[str]
pydantic-field
¶A list of tags for the item.
add_reference(self, ref_type, url, desc=None, force=False)
¶Source code in kiara/models/documentation.py
def add_reference(
self,
ref_type: str,
url: str,
desc: Optional[str] = None,
force: bool = False,
):
if ref_type in self.references.keys() and not force:
raise Exception(f"Reference of type '{ref_type}' already present.")
link = LinkModel(url=url, desc=desc)
self.references[ref_type] = link
create_renderable(self, **config)
¶Source code in kiara/models/documentation.py
def create_renderable(self, **config: Any) -> RenderableType:
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("Key", style="i")
table.add_column("Value")
if self.tags:
table.add_row("Tags", ", ".join(self.tags))
if self.labels:
labels = []
for k, v in self.labels.items():
labels.append(f"[i]{k}[/i]: {v}")
table.add_row("Labels", "\n".join(labels))
if self.references:
references = []
for _k, _v in self.references.items():
link = f"[link={_v.url}]{_v.url}[/link]"
references.append(f"[i]{_k}[/i]: {link}")
table.add_row("References", "\n".join(references))
return table
from_class(item_cls)
classmethod
¶Source code in kiara/models/documentation.py
@classmethod
def from_class(cls, item_cls: Type):
data = get_metadata_for_python_module_or_class(item_cls) # type: ignore
merged = merge_dicts(*data)
return cls.parse_obj(merged)
get_url_for_reference(self, ref)
¶Source code in kiara/models/documentation.py
def get_url_for_reference(self, ref: str) -> Optional[str]:
link = self.references.get(ref, None)
if not link:
return None
return link.url
DocumentationMetadataModel (KiaraModel)
pydantic-model
¶Source code in kiara/models/documentation.py
class DocumentationMetadataModel(KiaraModel):
_metadata_key = "documentation"
@classmethod
def from_class_doc(cls, item_cls: Type):
doc = item_cls.__doc__
if not doc:
doc = DEFAULT_NO_DESC_VALUE
doc = inspect.cleandoc(doc)
return cls.from_string(doc)
@classmethod
def from_function(cls, func: Callable):
doc = func.__doc__
if not doc:
doc = DEFAULT_NO_DESC_VALUE
doc = inspect.cleandoc(doc)
return cls.from_string(doc)
@classmethod
def from_string(cls, doc: Optional[str]):
if not doc:
doc = DEFAULT_NO_DESC_VALUE
if "\n" in doc:
desc, doc = doc.split("\n", maxsplit=1)
else:
desc = doc
doc = None
if doc:
doc = doc.strip()
return cls(description=desc.strip(), doc=doc)
@classmethod
def from_dict(cls, data: Mapping):
doc = data.get("doc", None)
desc = data.get("description", None)
if desc is None:
desc = data.get("desc", None)
if not doc and not desc:
return cls.from_string(DEFAULT_NO_DESC_VALUE)
elif doc and not desc:
return cls.from_string(doc)
elif desc and not doc:
return cls.from_string(desc)
else:
return DocumentationMetadataModel(description=desc, doc=doc)
@classmethod
def create(cls, item: Any):
if not item:
return cls.from_string(DEFAULT_NO_DESC_VALUE)
elif isinstance(item, DocumentationMetadataModel):
return item
elif isinstance(item, Mapping):
return cls.from_dict(item)
if isinstance(item, type):
return cls.from_class_doc(item)
elif isinstance(item, str):
return cls.from_string(item)
else:
raise TypeError(f"Can't create documentation from type '{type(item)}'.")
description: str = Field(
description="Short description of the item.", default=DEFAULT_NO_DESC_VALUE
)
doc: Optional[str] = Field(
description="Detailed documentation of the item (in markdown).", default=None
)
@property
def is_set(self) -> bool:
if self.description and self.description != DEFAULT_NO_DESC_VALUE:
return True
else:
return False
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return DOCUMENTATION_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.full_doc
@property
def full_doc(self):
if self.doc:
return f"{self.description}\n\n{self.doc}"
else:
return self.description
def create_renderable(self, **config: Any) -> RenderableType:
return Markdown(self.full_doc)
description: str
pydantic-field
¶Short description of the item.
doc: str
pydantic-field
¶Detailed documentation of the item (in markdown).
full_doc
property
readonly
¶is_set: bool
property
readonly
¶create(item)
classmethod
¶Source code in kiara/models/documentation.py
@classmethod
def create(cls, item: Any):
if not item:
return cls.from_string(DEFAULT_NO_DESC_VALUE)
elif isinstance(item, DocumentationMetadataModel):
return item
elif isinstance(item, Mapping):
return cls.from_dict(item)
if isinstance(item, type):
return cls.from_class_doc(item)
elif isinstance(item, str):
return cls.from_string(item)
else:
raise TypeError(f"Can't create documentation from type '{type(item)}'.")
create_renderable(self, **config)
¶Source code in kiara/models/documentation.py
def create_renderable(self, **config: Any) -> RenderableType:
return Markdown(self.full_doc)
from_class_doc(item_cls)
classmethod
¶Source code in kiara/models/documentation.py
@classmethod
def from_class_doc(cls, item_cls: Type):
doc = item_cls.__doc__
if not doc:
doc = DEFAULT_NO_DESC_VALUE
doc = inspect.cleandoc(doc)
return cls.from_string(doc)
from_dict(data)
classmethod
¶Source code in kiara/models/documentation.py
@classmethod
def from_dict(cls, data: Mapping):
doc = data.get("doc", None)
desc = data.get("description", None)
if desc is None:
desc = data.get("desc", None)
if not doc and not desc:
return cls.from_string(DEFAULT_NO_DESC_VALUE)
elif doc and not desc:
return cls.from_string(doc)
elif desc and not doc:
return cls.from_string(desc)
else:
return DocumentationMetadataModel(description=desc, doc=doc)
from_function(func)
classmethod
¶Source code in kiara/models/documentation.py
@classmethod
def from_function(cls, func: Callable):
doc = func.__doc__
if not doc:
doc = DEFAULT_NO_DESC_VALUE
doc = inspect.cleandoc(doc)
return cls.from_string(doc)
from_string(doc)
classmethod
¶Source code in kiara/models/documentation.py
@classmethod
def from_string(cls, doc: Optional[str]):
if not doc:
doc = DEFAULT_NO_DESC_VALUE
if "\n" in doc:
desc, doc = doc.split("\n", maxsplit=1)
else:
desc = doc
doc = None
if doc:
doc = doc.strip()
return cls(description=desc.strip(), doc=doc)
LinkModel (BaseModel)
pydantic-model
¶Source code in kiara/models/documentation.py
class LinkModel(BaseModel):
url: AnyUrl = Field(description="The url.")
desc: Optional[str] = Field(
description="A short description of the link content.",
default=DEFAULT_NO_DESC_VALUE,
)
events
special
¶
Classes¶
KiaraEvent (BaseModel)
pydantic-model
¶Source code in kiara/models/events/__init__.py
class KiaraEvent(BaseModel):
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
def get_event_type(self) -> str:
if hasattr(self, "event_type"):
return self.event_type # type: ignore
name = camel_case_to_snake_case(self.__class__.__name__)
return name
Config
¶Source code in kiara/models/events/__init__.py
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
json_loads
¶json_dumps(v, *, default=None, **args)
¶Source code in kiara/models/events/__init__.py
def orjson_dumps(v, *, default=None, **args):
# orjson.dumps returns bytes, to match standard json.dumps we need to decode
try:
return orjson.dumps(v, default=default, **args).decode()
except Exception as e:
if is_debug():
print(f"Error dumping json data: {e}")
from kiara import dbg
dbg(v)
raise e
get_event_type(self)
¶Source code in kiara/models/events/__init__.py
def get_event_type(self) -> str:
if hasattr(self, "event_type"):
return self.event_type # type: ignore
name = camel_case_to_snake_case(self.__class__.__name__)
return name
RegistryEvent (KiaraEvent)
pydantic-model
¶Modules¶
alias_registry
¶
AliasArchiveAddedEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/alias_registry.py
class AliasArchiveAddedEvent(RegistryEvent):
event_type: Literal["alias_archive_added"] = "alias_archive_added"
alias_archive_id: uuid.UUID = Field(
description="The unique id of this data archive."
)
alias_archive_alias: str = Field(
description="The alias this data archive was added as."
)
is_store: bool = Field(
description="Whether this archive supports write operations (aka implements the 'DataStore' interface)."
)
is_default_store: bool = Field(
description="Whether this store acts as default store."
)
alias_archive_alias: str
pydantic-field
required
¶The alias this data archive was added as.
alias_archive_id: UUID
pydantic-field
required
¶The unique id of this data archive.
event_type: Literal['alias_archive_added']
pydantic-field
¶is_default_store: bool
pydantic-field
required
¶Whether this store acts as default store.
is_store: bool
pydantic-field
required
¶Whether this archive supports write operations (aka implements the 'DataStore' interface).
AliasPreStoreEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/alias_registry.py
class AliasPreStoreEvent(RegistryEvent):
event_type: Literal["alias_pre_store"] = "alias_pre_store"
aliases: Iterable[str] = Field(description="The alias.")
AliasStoredEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/alias_registry.py
class AliasStoredEvent(RegistryEvent):
event_type: Literal["alias_stored"] = "alias_stored"
alias: str = Field(description="The alias.")
data_registry
¶
DataArchiveAddedEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/data_registry.py
class DataArchiveAddedEvent(RegistryEvent):
event_type: Literal["data_archive_added"] = "data_archive_added"
data_archive_id: uuid.UUID = Field(
description="The unique id of this data archive."
)
data_archive_alias: str = Field(
description="The alias this data archive was added as."
)
is_store: bool = Field(
description="Whether this archive supports write operations (aka implements the 'DataStore' interface)."
)
is_default_store: bool = Field(
description="Whether this store acts as default store."
)
data_archive_alias: str
pydantic-field
required
¶The alias this data archive was added as.
data_archive_id: UUID
pydantic-field
required
¶The unique id of this data archive.
event_type: Literal['data_archive_added']
pydantic-field
¶is_default_store: bool
pydantic-field
required
¶Whether this store acts as default store.
is_store: bool
pydantic-field
required
¶Whether this archive supports write operations (aka implements the 'DataStore' interface).
ValueCreatedEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/data_registry.py
class ValueCreatedEvent(RegistryEvent):
event_type: Literal["value_created"] = "value_created"
value: Value = Field(description="The value metadata.")
ValuePreStoreEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/data_registry.py
class ValuePreStoreEvent(RegistryEvent):
event_type: Literal["value_pre_store"] = "value_pre_store"
value: Value = Field(description="The value metadata.")
ValueStoredEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/data_registry.py
class ValueStoredEvent(RegistryEvent):
event_type: Literal["value_stored"] = "value_stored"
value: Value = Field(description="The value metadata.")
job_registry
¶
JobArchiveAddedEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/job_registry.py
class JobArchiveAddedEvent(RegistryEvent):
event_type: Literal["job_archive_added"] = "job_archive_added"
job_archive_id: uuid.UUID = Field(description="The unique id of this job archive.")
job_archive_alias: str = Field(
description="The alias this job archive was added as."
)
is_store: bool = Field(
description="Whether this archive supports write operations (aka implements the 'JobStore' interface)."
)
is_default_store: bool = Field(
description="Whether this store acts as default store."
)
event_type: Literal['job_archive_added']
pydantic-field
¶is_default_store: bool
pydantic-field
required
¶Whether this store acts as default store.
is_store: bool
pydantic-field
required
¶Whether this archive supports write operations (aka implements the 'JobStore' interface).
job_archive_alias: str
pydantic-field
required
¶The alias this job archive was added as.
job_archive_id: UUID
pydantic-field
required
¶The unique id of this job archive.
JobRecordPreStoreEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/job_registry.py
class JobRecordPreStoreEvent(RegistryEvent):
event_type: Literal["job_record_pre_store"] = "job_record_pre_store"
job_record: JobRecord = Field(description="The job record.")
JobRecordStoredEvent (RegistryEvent)
pydantic-model
¶Source code in kiara/models/events/job_registry.py
class JobRecordStoredEvent(RegistryEvent):
event_type: Literal["job_record_stored"] = "job_record_stored"
job_record: JobRecord = Field(description="The job record.")
pipeline
¶
ChangedValue (BaseModel)
pydantic-model
¶
PipelineDetails (BaseModel)
pydantic-model
¶Source code in kiara/models/events/pipeline.py
class PipelineDetails(BaseModel):
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
kiara_id: uuid.UUID = Field(description="The id of the kiara context.")
pipeline_id: uuid.UUID = Field(description="The id of the pipeline.")
pipeline_status: StepStatus = Field(
description="The current status of this pipeline."
)
invalid_details: Dict[str, str] = Field(
description="Details about fields that are invalid (if status < 'INPUTS_READY'.",
default_factory=dict,
)
pipeline_inputs: Dict[str, uuid.UUID] = Field(
description="The current pipeline inputs."
)
pipeline_outputs: Dict[str, uuid.UUID] = Field(
description="The current pipeline outputs."
)
step_states: Dict[str, StepDetails] = Field(
description="The state of each step within this pipeline."
)
def get_steps_by_processing_stage(self) -> MutableMapping[int, List[StepDetails]]:
result: MutableMapping[int, List[StepDetails]] = SortedDict()
for step_details in self.step_states.values():
result.setdefault(step_details.processing_stage, []).append(step_details)
return result
invalid_details: Dict[str, str]
pydantic-field
¶Details about fields that are invalid (if status < 'INPUTS_READY'.
kiara_id: UUID
pydantic-field
required
¶The id of the kiara context.
pipeline_id: UUID
pydantic-field
required
¶The id of the pipeline.
pipeline_inputs: Dict[str, uuid.UUID]
pydantic-field
required
¶The current pipeline inputs.
pipeline_outputs: Dict[str, uuid.UUID]
pydantic-field
required
¶The current pipeline outputs.
pipeline_status: StepStatus
pydantic-field
required
¶The current status of this pipeline.
step_states: Dict[str, kiara.models.events.pipeline.StepDetails]
pydantic-field
required
¶The state of each step within this pipeline.
Config
¶Source code in kiara/models/events/pipeline.py
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
json_loads
¶json_dumps(v, *, default=None, **args)
¶Source code in kiara/models/events/pipeline.py
def orjson_dumps(v, *, default=None, **args):
# orjson.dumps returns bytes, to match standard json.dumps we need to decode
try:
return orjson.dumps(v, default=default, **args).decode()
except Exception as e:
if is_debug():
print(f"Error dumping json data: {e}")
from kiara import dbg
dbg(v)
raise e
get_steps_by_processing_stage(self)
¶Source code in kiara/models/events/pipeline.py
def get_steps_by_processing_stage(self) -> MutableMapping[int, List[StepDetails]]:
result: MutableMapping[int, List[StepDetails]] = SortedDict()
for step_details in self.step_states.values():
result.setdefault(step_details.processing_stage, []).append(step_details)
return result
PipelineEvent (KiaraEvent)
pydantic-model
¶Source code in kiara/models/events/pipeline.py
class PipelineEvent(KiaraEvent):
@classmethod
def create_event(
cls,
pipeline: "Pipeline",
changed: Mapping[str, Mapping[str, Mapping[str, ChangedValue]]],
):
pipeline_inputs = changed.get("__pipeline__", {}).get("inputs", {})
pipeline_outputs = changed.get("__pipeline__", {}).get("outputs", {})
step_inputs = {}
step_outputs = {}
invalidated_steps: Set[str] = set()
for step_id, change_details in changed.items():
if step_id == "__pipeline__":
continue
inputs = change_details.get("inputs", None)
if inputs:
invalidated_steps.add(step_id)
step_inputs[step_id] = inputs
outputs = change_details.get("outputs", None)
if outputs:
invalidated_steps.add(step_id)
step_outputs[step_id] = outputs
event = PipelineEvent(
kiara_id=pipeline.kiara_id,
pipeline_id=pipeline.pipeline_id,
pipeline_inputs_changed=pipeline_inputs,
pipeline_outputs_changed=pipeline_outputs,
step_inputs_changed=step_inputs,
step_outputs_changed=step_outputs,
changed_steps=sorted(invalidated_steps),
)
return event
class Config:
allow_mutation = False
kiara_id: uuid.UUID = Field(
description="The id of the kiara context that created the pipeline."
)
pipeline_id: uuid.UUID = Field(description="The pipeline id.")
pipeline_inputs_changed: Dict[str, ChangedValue] = Field(
description="Details about changed pipeline input values.", default_factory=dict
)
pipeline_outputs_changed: Dict[str, ChangedValue] = Field(
description="Details about changed pipeline output values.",
default_factory=dict,
)
step_inputs_changed: Dict[str, Mapping[str, ChangedValue]] = Field(
description="Details about changed step input values.", default_factory=dict
)
step_outputs_changed: Dict[str, Mapping[str, ChangedValue]] = Field(
description="Details about changed step output values.", default_factory=dict
)
changed_steps: List[str] = Field(
description="A list of all step ids that have newly invalidated outputs."
)
def __repr__(self):
return f"{self.__class__.__name__}(pipeline_id={self.pipeline_id}, invalidated_steps={', '.join(self.changed_steps)})"
def __str__(self):
return self.__repr__()
changed_steps: List[str]
pydantic-field
required
¶A list of all step ids that have newly invalidated outputs.
kiara_id: UUID
pydantic-field
required
¶The id of the kiara context that created the pipeline.
pipeline_id: UUID
pydantic-field
required
¶The pipeline id.
pipeline_inputs_changed: Dict[str, kiara.models.events.pipeline.ChangedValue]
pydantic-field
¶Details about changed pipeline input values.
pipeline_outputs_changed: Dict[str, kiara.models.events.pipeline.ChangedValue]
pydantic-field
¶Details about changed pipeline output values.
step_inputs_changed: Dict[str, Mapping[str, kiara.models.events.pipeline.ChangedValue]]
pydantic-field
¶Details about changed step input values.
step_outputs_changed: Dict[str, Mapping[str, kiara.models.events.pipeline.ChangedValue]]
pydantic-field
¶Details about changed step output values.
Config
¶Source code in kiara/models/events/pipeline.py
class Config:
allow_mutation = False
create_event(pipeline, changed)
classmethod
¶Source code in kiara/models/events/pipeline.py
@classmethod
def create_event(
cls,
pipeline: "Pipeline",
changed: Mapping[str, Mapping[str, Mapping[str, ChangedValue]]],
):
pipeline_inputs = changed.get("__pipeline__", {}).get("inputs", {})
pipeline_outputs = changed.get("__pipeline__", {}).get("outputs", {})
step_inputs = {}
step_outputs = {}
invalidated_steps: Set[str] = set()
for step_id, change_details in changed.items():
if step_id == "__pipeline__":
continue
inputs = change_details.get("inputs", None)
if inputs:
invalidated_steps.add(step_id)
step_inputs[step_id] = inputs
outputs = change_details.get("outputs", None)
if outputs:
invalidated_steps.add(step_id)
step_outputs[step_id] = outputs
event = PipelineEvent(
kiara_id=pipeline.kiara_id,
pipeline_id=pipeline.pipeline_id,
pipeline_inputs_changed=pipeline_inputs,
pipeline_outputs_changed=pipeline_outputs,
step_inputs_changed=step_inputs,
step_outputs_changed=step_outputs,
changed_steps=sorted(invalidated_steps),
)
return event
StepDetails (BaseModel)
pydantic-model
¶Source code in kiara/models/events/pipeline.py
class StepDetails(BaseModel):
kiara_id: uuid.UUID = Field(description="The id of the kiara context.")
pipeline_id: uuid.UUID = Field(description="The id of the pipeline.")
step_id: str = Field(description="The id of the step.")
processing_stage: int = Field(
description="The execution stage where this step is executed."
)
status: StepStatus = Field(description="The current status of this step.")
invalid_details: Dict[str, str] = Field(
description="Details about fields that are invalid (if status < 'INPUTS_READY'.",
default_factory=dict,
)
inputs: Dict[str, uuid.UUID] = Field(description="The current inputs of this step.")
outputs: Dict[str, uuid.UUID] = Field(
description="The current outputs of this step."
)
@validator("inputs")
def replace_none_values_inputs(cls, value):
result = {}
for k, v in value.items():
if v is None:
v = NONE_VALUE_ID
result[k] = v
return result
@validator("outputs")
def replace_none_values_outputs(cls, value):
result = {}
for k, v in value.items():
if v is None:
v = NOT_SET_VALUE_ID
result[k] = v
return result
def _retrieve_data_to_hash(self) -> Any:
return {
"kiara_id": self.kiara_id,
"pipeline_id": self.pipeline_id,
"step_id": self.step_id,
"inputs": self.inputs,
}
def _retrieve_id(self) -> str:
return f"{self.kiara_id}.{self.pipeline_id}.{self.step_id}"
def _retrieve_category_id(self) -> str:
return PIPELINE_STEP_DETAILS_CATEGORY_ID
inputs: Dict[str, uuid.UUID]
pydantic-field
required
¶The current inputs of this step.
invalid_details: Dict[str, str]
pydantic-field
¶Details about fields that are invalid (if status < 'INPUTS_READY'.
kiara_id: UUID
pydantic-field
required
¶The id of the kiara context.
outputs: Dict[str, uuid.UUID]
pydantic-field
required
¶The current outputs of this step.
pipeline_id: UUID
pydantic-field
required
¶The id of the pipeline.
processing_stage: int
pydantic-field
required
¶The execution stage where this step is executed.
status: StepStatus
pydantic-field
required
¶The current status of this step.
step_id: str
pydantic-field
required
¶The id of the step.
replace_none_values_inputs(value)
classmethod
¶Source code in kiara/models/events/pipeline.py
@validator("inputs")
def replace_none_values_inputs(cls, value):
result = {}
for k, v in value.items():
if v is None:
v = NONE_VALUE_ID
result[k] = v
return result
replace_none_values_outputs(value)
classmethod
¶Source code in kiara/models/events/pipeline.py
@validator("outputs")
def replace_none_values_outputs(cls, value):
result = {}
for k, v in value.items():
if v is None:
v = NOT_SET_VALUE_ID
result[k] = v
return result
filesystem
¶
logger
¶Classes¶
FileBundle (KiaraModel)
pydantic-model
¶Describes properties for the 'file_bundle' value type.
Source code in kiara/models/filesystem.py
class FileBundle(KiaraModel):
"""Describes properties for the 'file_bundle' value type."""
@classmethod
def import_folder(
cls,
source: str,
bundle_name: Optional[str] = None,
import_config: Union[None, Mapping[str, Any], FolderImportConfig] = None,
import_time: Optional[datetime.datetime] = None,
) -> "FileBundle":
if not source:
raise ValueError("No source path provided.")
if not os.path.exists(os.path.realpath(source)):
raise ValueError(f"Path does not exist: {source}")
if not os.path.isdir(os.path.realpath(source)):
raise ValueError(f"Path is not a file: {source}")
if source.endswith(os.path.sep):
source = source[0:-1]
abs_path = os.path.abspath(source)
if import_config is None:
_import_config = FolderImportConfig()
elif isinstance(import_config, Mapping):
_import_config = FolderImportConfig(**import_config)
elif isinstance(import_config, FolderImportConfig):
_import_config = import_config
else:
raise TypeError(
f"Invalid type for folder import config: {type(import_config)}."
)
included_files: Dict[str, FileModel] = {}
exclude_dirs = _import_config.exclude_dirs
invalid_extensions = _import_config.exclude_files
valid_extensions = _import_config.include_files
if import_time:
bundle_import_time = import_time
else:
bundle_import_time = datetime.datetime.now() # TODO: timezone
sum_size = 0
def include_file(filename: str) -> bool:
if invalid_extensions and any(
filename.endswith(ext) for ext in invalid_extensions
):
return False
if not valid_extensions:
return True
else:
return any(filename.endswith(ext) for ext in valid_extensions)
for root, dirnames, filenames in os.walk(abs_path, topdown=True):
if exclude_dirs:
dirnames[:] = [d for d in dirnames if d not in exclude_dirs]
for filename in [
f
for f in filenames
if os.path.isfile(os.path.join(root, f)) and include_file(f)
]:
full_path = os.path.join(root, filename)
rel_path = os.path.relpath(full_path, abs_path)
file_model = FileModel.load_file(
full_path, import_time=bundle_import_time
)
sum_size = sum_size + file_model.size
included_files[rel_path] = file_model
if bundle_name is None:
bundle_name = os.path.basename(source)
bundle = FileBundle.create_from_file_models(
files=included_files,
path=abs_path,
bundle_name=bundle_name,
sum_size=sum_size,
import_time=bundle_import_time,
)
return bundle
@classmethod
def create_from_file_models(
cls,
files: Mapping[str, FileModel],
bundle_name: str,
path: str,
sum_size: Optional[int] = None,
import_time: Optional[datetime.datetime] = None,
) -> "FileBundle":
if import_time:
bundle_import_time = import_time
else:
bundle_import_time = datetime.datetime.now() # TODO: timezone
result: Dict[str, Any] = {}
result["included_files"] = files
result["import_time"] = datetime.datetime.now().isoformat()
result["number_of_files"] = len(files)
result["bundle_name"] = bundle_name
result["import_time"] = bundle_import_time
if sum_size is None:
sum_size = 0
for f in files.values():
sum_size = sum_size + f.size
result["size"] = sum_size
bundle = FileBundle(**result)
bundle._path = path
return bundle
_file_bundle_hash: Optional[int] = PrivateAttr(default=None)
bundle_name: str = Field(description="The name of this bundle.")
import_time: datetime.datetime = Field(
description="The time when the file bundle was imported."
)
number_of_files: int = Field(
description="How many files are included in this bundle."
)
included_files: Dict[str, FileModel] = Field(
description="A map of all the included files, incl. their properties. Uses the relative path of each file as key."
)
size: int = Field(description="The size of all files in this folder, combined.")
_path: Optional[str] = PrivateAttr(default=None)
@property
def path(self) -> str:
if self._path is None:
raise Exception("File bundle path not set.")
return self._path
def _retrieve_id(self) -> str:
return str(self.file_bundle_hash)
def _retrieve_category_id(self) -> str:
return FILE_BUNDLE_MODEL_CATEOGORY_ID
@property
def model_data_hash(self) -> int:
return self.file_bundle_hash
def _retrieve_data_to_hash(self) -> Any:
raise NotImplementedError()
def get_relative_path(self, file: FileModel):
return os.path.relpath(file.path, self.path)
def read_text_file_contents(self, ignore_errors: bool = False) -> Mapping[str, str]:
content_dict: Dict[str, str] = {}
def read_file(rel_path: str, fm: FileModel):
with open(fm.path, encoding="utf-8") as f:
try:
content = f.read()
content_dict[rel_path] = content # type: ignore
except Exception as e:
if ignore_errors:
log_message(f"Can't read file: {e}")
logger.warning("ignore.file", path=fm.path, reason=str(e))
else:
raise Exception(f"Can't read file (as text) '{fm.path}: {e}")
# TODO: common ignore files and folders
for f in self.included_files.values():
rel_path = self.get_relative_path(f)
read_file(rel_path=rel_path, fm=f)
return content_dict
@property
def file_bundle_hash(self) -> int:
# TODO: use sha256?
if self._file_bundle_hash is not None:
return self._file_bundle_hash
obj = {k: v.file_hash for k, v in self.included_files.items()}
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
self._file_bundle_hash = h[obj]
return self._file_bundle_hash
def copy_bundle(
self, target_path: str, bundle_name: Optional[str] = None
) -> "FileBundle":
if target_path == self.path:
raise Exception(f"Target path and current path are the same: {target_path}")
result = {}
for rel_path, item in self.included_files.items():
_target_path = os.path.join(target_path, rel_path)
new_fm = item.copy_file(_target_path)
result[rel_path] = new_fm
if bundle_name is None:
bundle_name = os.path.basename(target_path)
fb = FileBundle.create_from_file_models(
files=result,
bundle_name=bundle_name,
path=target_path,
sum_size=self.size,
import_time=self.import_time,
)
if self._file_bundle_hash is not None:
fb._file_bundle_hash = self._file_bundle_hash
return fb
def create_renderable(self, **config: Any) -> RenderableType:
show_bundle_hash = config.get("show_bundle_hash", False)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("key")
table.add_column("value", style="i")
table.add_row("bundle name", self.bundle_name)
table.add_row("import_time", str(self.import_time))
table.add_row("number_of_files", str(self.number_of_files))
table.add_row("size", str(self.size))
if show_bundle_hash:
table.add_row("bundle_hash", str(self.file_bundle_hash))
content = self._create_content_table(**config)
table.add_row("included files", content)
return table
def _create_content_table(self, **render_config: Any) -> Table:
# show_content = render_config.get("show_content_preview", False)
max_no_included_files = render_config.get("max_no_files", 40)
table = Table(show_header=True, box=box.SIMPLE)
table.add_column("(relative) path")
table.add_column("size")
# if show_content:
# table.add_column("content preview")
if (
max_no_included_files < 0
or len(self.included_files) <= max_no_included_files
):
for f, model in self.included_files.items():
row = [f, str(model.size)]
table.add_row(*row)
else:
files = list(self.included_files.keys())
half = int((max_no_included_files - 1) / 2)
head = files[0:half]
tail = files[-1 * half :] # noqa
for rel_path in head:
model = self.included_files[rel_path]
row = [rel_path, str(model.size)]
table.add_row(*row)
table.add_row(" ... output skipped ...", "")
table.add_row(" ... output skipped ...", "")
for rel_path in tail:
model = self.included_files[rel_path]
row = [rel_path, str(model.size)]
table.add_row(*row)
return table
def __repr__(self):
return f"FileBundle(name={self.bundle_name})"
def __str__(self):
return self.__repr__()
bundle_name: str
pydantic-field
required
¶The name of this bundle.
file_bundle_hash: int
property
readonly
¶import_time: datetime
pydantic-field
required
¶The time when the file bundle was imported.
included_files: Dict[str, kiara.models.filesystem.FileModel]
pydantic-field
required
¶A map of all the included files, incl. their properties. Uses the relative path of each file as key.
model_data_hash: int
property
readonly
¶A hash for this model.
number_of_files: int
pydantic-field
required
¶How many files are included in this bundle.
path: str
property
readonly
¶size: int
pydantic-field
required
¶The size of all files in this folder, combined.
copy_bundle(self, target_path, bundle_name=None)
¶Source code in kiara/models/filesystem.py
def copy_bundle(
self, target_path: str, bundle_name: Optional[str] = None
) -> "FileBundle":
if target_path == self.path:
raise Exception(f"Target path and current path are the same: {target_path}")
result = {}
for rel_path, item in self.included_files.items():
_target_path = os.path.join(target_path, rel_path)
new_fm = item.copy_file(_target_path)
result[rel_path] = new_fm
if bundle_name is None:
bundle_name = os.path.basename(target_path)
fb = FileBundle.create_from_file_models(
files=result,
bundle_name=bundle_name,
path=target_path,
sum_size=self.size,
import_time=self.import_time,
)
if self._file_bundle_hash is not None:
fb._file_bundle_hash = self._file_bundle_hash
return fb
create_from_file_models(files, bundle_name, path, sum_size=None, import_time=None)
classmethod
¶Source code in kiara/models/filesystem.py
@classmethod
def create_from_file_models(
cls,
files: Mapping[str, FileModel],
bundle_name: str,
path: str,
sum_size: Optional[int] = None,
import_time: Optional[datetime.datetime] = None,
) -> "FileBundle":
if import_time:
bundle_import_time = import_time
else:
bundle_import_time = datetime.datetime.now() # TODO: timezone
result: Dict[str, Any] = {}
result["included_files"] = files
result["import_time"] = datetime.datetime.now().isoformat()
result["number_of_files"] = len(files)
result["bundle_name"] = bundle_name
result["import_time"] = bundle_import_time
if sum_size is None:
sum_size = 0
for f in files.values():
sum_size = sum_size + f.size
result["size"] = sum_size
bundle = FileBundle(**result)
bundle._path = path
return bundle
create_renderable(self, **config)
¶Source code in kiara/models/filesystem.py
def create_renderable(self, **config: Any) -> RenderableType:
show_bundle_hash = config.get("show_bundle_hash", False)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("key")
table.add_column("value", style="i")
table.add_row("bundle name", self.bundle_name)
table.add_row("import_time", str(self.import_time))
table.add_row("number_of_files", str(self.number_of_files))
table.add_row("size", str(self.size))
if show_bundle_hash:
table.add_row("bundle_hash", str(self.file_bundle_hash))
content = self._create_content_table(**config)
table.add_row("included files", content)
return table
get_relative_path(self, file)
¶Source code in kiara/models/filesystem.py
def get_relative_path(self, file: FileModel):
return os.path.relpath(file.path, self.path)
import_folder(source, bundle_name=None, import_config=None, import_time=None)
classmethod
¶Source code in kiara/models/filesystem.py
@classmethod
def import_folder(
cls,
source: str,
bundle_name: Optional[str] = None,
import_config: Union[None, Mapping[str, Any], FolderImportConfig] = None,
import_time: Optional[datetime.datetime] = None,
) -> "FileBundle":
if not source:
raise ValueError("No source path provided.")
if not os.path.exists(os.path.realpath(source)):
raise ValueError(f"Path does not exist: {source}")
if not os.path.isdir(os.path.realpath(source)):
raise ValueError(f"Path is not a file: {source}")
if source.endswith(os.path.sep):
source = source[0:-1]
abs_path = os.path.abspath(source)
if import_config is None:
_import_config = FolderImportConfig()
elif isinstance(import_config, Mapping):
_import_config = FolderImportConfig(**import_config)
elif isinstance(import_config, FolderImportConfig):
_import_config = import_config
else:
raise TypeError(
f"Invalid type for folder import config: {type(import_config)}."
)
included_files: Dict[str, FileModel] = {}
exclude_dirs = _import_config.exclude_dirs
invalid_extensions = _import_config.exclude_files
valid_extensions = _import_config.include_files
if import_time:
bundle_import_time = import_time
else:
bundle_import_time = datetime.datetime.now() # TODO: timezone
sum_size = 0
def include_file(filename: str) -> bool:
if invalid_extensions and any(
filename.endswith(ext) for ext in invalid_extensions
):
return False
if not valid_extensions:
return True
else:
return any(filename.endswith(ext) for ext in valid_extensions)
for root, dirnames, filenames in os.walk(abs_path, topdown=True):
if exclude_dirs:
dirnames[:] = [d for d in dirnames if d not in exclude_dirs]
for filename in [
f
for f in filenames
if os.path.isfile(os.path.join(root, f)) and include_file(f)
]:
full_path = os.path.join(root, filename)
rel_path = os.path.relpath(full_path, abs_path)
file_model = FileModel.load_file(
full_path, import_time=bundle_import_time
)
sum_size = sum_size + file_model.size
included_files[rel_path] = file_model
if bundle_name is None:
bundle_name = os.path.basename(source)
bundle = FileBundle.create_from_file_models(
files=included_files,
path=abs_path,
bundle_name=bundle_name,
sum_size=sum_size,
import_time=bundle_import_time,
)
return bundle
read_text_file_contents(self, ignore_errors=False)
¶Source code in kiara/models/filesystem.py
def read_text_file_contents(self, ignore_errors: bool = False) -> Mapping[str, str]:
content_dict: Dict[str, str] = {}
def read_file(rel_path: str, fm: FileModel):
with open(fm.path, encoding="utf-8") as f:
try:
content = f.read()
content_dict[rel_path] = content # type: ignore
except Exception as e:
if ignore_errors:
log_message(f"Can't read file: {e}")
logger.warning("ignore.file", path=fm.path, reason=str(e))
else:
raise Exception(f"Can't read file (as text) '{fm.path}: {e}")
# TODO: common ignore files and folders
for f in self.included_files.values():
rel_path = self.get_relative_path(f)
read_file(rel_path=rel_path, fm=f)
return content_dict
FileModel (KiaraModel)
pydantic-model
¶Describes properties for the 'file' value type.
Source code in kiara/models/filesystem.py
class FileModel(KiaraModel):
"""Describes properties for the 'file' value type."""
@classmethod
def load_file(
cls,
source: str,
file_name: Optional[str] = None,
import_time: Optional[datetime.datetime] = None,
):
"""Utility method to read metadata of a file from disk and optionally move it into a data archive location."""
import filetype
import mimetypes
if not source:
raise ValueError("No source path provided.")
if not os.path.exists(os.path.realpath(source)):
raise ValueError(f"Path does not exist: {source}")
if not os.path.isfile(os.path.realpath(source)):
raise ValueError(f"Path is not a file: {source}")
if file_name is None:
file_name = os.path.basename(source)
path: str = os.path.abspath(source)
if import_time:
file_import_time = import_time
else:
file_import_time = datetime.datetime.now() # TODO: timezone
file_stats = os.stat(path)
size = file_stats.st_size
r = mimetypes.guess_type(path)
if r[0] is not None:
mime_type = r[0]
else:
_mime_type = filetype.guess(path)
if not _mime_type:
mime_type = "application/octet-stream"
else:
mime_type = _mime_type.MIME
m = FileModel(
import_time=file_import_time,
mime_type=mime_type,
size=size,
file_name=file_name,
)
m._path = path
return m
_file_hash: Optional[str] = PrivateAttr(default=None)
import_time: datetime.datetime = Field(
description="The time when the file was imported."
)
mime_type: str = Field(description="The mime type of the file.")
file_name: str = Field("The name of the file.")
size: int = Field(description="The size of the file.")
_path: Optional[str] = PrivateAttr(default=None)
# @validator("path")
# def ensure_abs_path(cls, value):
# return os.path.abspath(value)
@property
def path(self) -> str:
if self._path is None:
raise Exception("File path not set for file model.")
return self._path
def _retrieve_id(self) -> str:
return self.file_hash
def _retrieve_category_id(self) -> str:
return FILE_MODEL_CATEOGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return {
"file_name": self.file_name,
"file_hash": self.file_hash,
}
# def get_id(self) -> str:
# return self.path
def get_category_alias(self) -> str:
return "instance.file_model"
def copy_file(self, target: str, new_name: Optional[str] = None) -> "FileModel":
target_path: str = os.path.abspath(target)
os.makedirs(os.path.dirname(target_path), exist_ok=True)
shutil.copy2(self.path, target_path)
fm = FileModel.load_file(
target, file_name=new_name, import_time=self.import_time
)
if self._file_hash is not None:
fm._file_hash = self._file_hash
return fm
@property
def file_hash(self) -> str:
if self._file_hash is not None:
return self._file_hash
sha256_hash = hashlib.sha3_256()
with open(self.path, "rb") as f:
# Read and update hash string value in blocks of 4K
for byte_block in iter(lambda: f.read(4096), b""):
sha256_hash.update(byte_block)
self._file_hash = sha256_hash.hexdigest()
return self._file_hash
@property
def file_name_without_extension(self) -> str:
return self.file_name.split(".")[0]
def read_text(self, max_lines: int = -1) -> str:
"""Read the content of a file."""
with open(self.path, "rt") as f:
if max_lines <= 0:
content = f.read()
else:
content = "".join((next(f) for x in range(max_lines)))
return content
def read_bytes(self, length: int = -1) -> bytes:
"""Read the content of a file."""
with open(self.path, "rb") as f:
if length <= 0:
content = f.read()
else:
content = f.read(length)
return content
def __repr__(self):
return f"FileModel(name={self.file_name})"
def __str__(self):
return self.__repr__()
file_hash: str
property
readonly
¶file_name: str
pydantic-field
¶file_name_without_extension: str
property
readonly
¶import_time: datetime
pydantic-field
required
¶The time when the file was imported.
mime_type: str
pydantic-field
required
¶The mime type of the file.
path: str
property
readonly
¶size: int
pydantic-field
required
¶The size of the file.
copy_file(self, target, new_name=None)
¶Source code in kiara/models/filesystem.py
def copy_file(self, target: str, new_name: Optional[str] = None) -> "FileModel":
target_path: str = os.path.abspath(target)
os.makedirs(os.path.dirname(target_path), exist_ok=True)
shutil.copy2(self.path, target_path)
fm = FileModel.load_file(
target, file_name=new_name, import_time=self.import_time
)
if self._file_hash is not None:
fm._file_hash = self._file_hash
return fm
get_category_alias(self)
¶Source code in kiara/models/filesystem.py
def get_category_alias(self) -> str:
return "instance.file_model"
load_file(source, file_name=None, import_time=None)
classmethod
¶Utility method to read metadata of a file from disk and optionally move it into a data archive location.
Source code in kiara/models/filesystem.py
@classmethod
def load_file(
cls,
source: str,
file_name: Optional[str] = None,
import_time: Optional[datetime.datetime] = None,
):
"""Utility method to read metadata of a file from disk and optionally move it into a data archive location."""
import filetype
import mimetypes
if not source:
raise ValueError("No source path provided.")
if not os.path.exists(os.path.realpath(source)):
raise ValueError(f"Path does not exist: {source}")
if not os.path.isfile(os.path.realpath(source)):
raise ValueError(f"Path is not a file: {source}")
if file_name is None:
file_name = os.path.basename(source)
path: str = os.path.abspath(source)
if import_time:
file_import_time = import_time
else:
file_import_time = datetime.datetime.now() # TODO: timezone
file_stats = os.stat(path)
size = file_stats.st_size
r = mimetypes.guess_type(path)
if r[0] is not None:
mime_type = r[0]
else:
_mime_type = filetype.guess(path)
if not _mime_type:
mime_type = "application/octet-stream"
else:
mime_type = _mime_type.MIME
m = FileModel(
import_time=file_import_time,
mime_type=mime_type,
size=size,
file_name=file_name,
)
m._path = path
return m
read_bytes(self, length=-1)
¶Read the content of a file.
Source code in kiara/models/filesystem.py
def read_bytes(self, length: int = -1) -> bytes:
"""Read the content of a file."""
with open(self.path, "rb") as f:
if length <= 0:
content = f.read()
else:
content = f.read(length)
return content
read_text(self, max_lines=-1)
¶Read the content of a file.
Source code in kiara/models/filesystem.py
def read_text(self, max_lines: int = -1) -> str:
"""Read the content of a file."""
with open(self.path, "rt") as f:
if max_lines <= 0:
content = f.read()
else:
content = "".join((next(f) for x in range(max_lines)))
return content
FolderImportConfig (BaseModel)
pydantic-model
¶Source code in kiara/models/filesystem.py
class FolderImportConfig(BaseModel):
include_files: Optional[List[str]] = Field(
description="A list of strings, include all files where the filename ends with that string.",
default=None,
)
exclude_dirs: Optional[List[str]] = Field(
description="A list of strings, exclude all folders whose name ends with that string.",
default=None,
)
exclude_files: Optional[List[str]] = Field(
description=f"A list of strings, exclude all files that match those (takes precedence over 'include_files'). Defaults to: {DEFAULT_EXCLUDE_FILES}.",
default=DEFAULT_EXCLUDE_FILES,
)
exclude_dirs: List[str]
pydantic-field
¶A list of strings, exclude all folders whose name ends with that string.
exclude_files: List[str]
pydantic-field
¶A list of strings, exclude all files that match those (takes precedence over 'include_files'). Defaults to: ['.DS_Store'].
include_files: List[str]
pydantic-field
¶A list of strings, include all files where the filename ends with that string.
info
¶
INFO_BASE_CLASS
¶INFO_CLASS
¶Classes¶
InfoModelGroup (KiaraModel, Mapping, Generic)
pydantic-model
¶Source code in kiara/models/info.py
class InfoModelGroup(KiaraModel, Mapping[str, KiaraInfoModel]):
@classmethod
@abc.abstractmethod
def base_info_class(cls) -> Type[KiaraInfoModel]:
pass
group_id: uuid.UUID = Field(
description="The unique group id.", default_factory=ID_REGISTRY.generate
)
group_alias: Optional[str] = Field(description="The group alias.", default=None)
def _retrieve_id(self) -> str:
return str(self.group_id)
def _retrieve_category_id(self) -> str:
return f"type_info_group.{self.type_name}" # type: ignore
def _retrieve_subcomponent_keys(self) -> Iterable[str]:
return self.type_infos.keys() # type: ignore
def _retrieve_data_to_hash(self) -> Any:
return {"type_name": self.type_name, "included_types": self.type_infos.keys()} # type: ignore
def get_type_infos(self) -> Mapping[str, KiaraInfoModel]:
return self.type_infos # type: ignore
def create_renderable(self, **config: Any) -> RenderableType:
full_doc = config.get("full_doc", False)
table = Table(show_header=True, box=box.SIMPLE, show_lines=full_doc)
table.add_column("Type name", style="i")
table.add_column("Description")
for type_name in sorted(self.type_infos.keys()): # type: ignore
t_md = self.type_infos[type_name] # type: ignore
if full_doc:
md = Markdown(t_md.documentation.full_doc)
else:
md = Markdown(t_md.documentation.description)
table.add_row(type_name, md)
return table
def __getitem__(self, item: str) -> KiaraInfoModel:
return self.get_type_infos()[item]
def __iter__(self):
return iter(self.get_type_infos())
def __len__(self):
return len(self.get_type_infos())
group_alias: str
pydantic-field
¶The group alias.
group_id: UUID
pydantic-field
¶The unique group id.
base_info_class()
classmethod
¶Source code in kiara/models/info.py
@classmethod
@abc.abstractmethod
def base_info_class(cls) -> Type[KiaraInfoModel]:
pass
create_renderable(self, **config)
¶Source code in kiara/models/info.py
def create_renderable(self, **config: Any) -> RenderableType:
full_doc = config.get("full_doc", False)
table = Table(show_header=True, box=box.SIMPLE, show_lines=full_doc)
table.add_column("Type name", style="i")
table.add_column("Description")
for type_name in sorted(self.type_infos.keys()): # type: ignore
t_md = self.type_infos[type_name] # type: ignore
if full_doc:
md = Markdown(t_md.documentation.full_doc)
else:
md = Markdown(t_md.documentation.description)
table.add_row(type_name, md)
return table
get_type_infos(self)
¶Source code in kiara/models/info.py
def get_type_infos(self) -> Mapping[str, KiaraInfoModel]:
return self.type_infos # type: ignore
KiaraInfoModel (KiaraModel, Generic)
pydantic-model
¶Source code in kiara/models/info.py
class KiaraInfoModel(KiaraModel, Generic[INFO_BASE_CLASS]):
@classmethod
@abc.abstractmethod
def category_name(cls) -> str:
pass
@validator("documentation", pre=True)
def validate_doc(cls, value):
return DocumentationMetadataModel.create(value)
type_name: str = Field(description="The registered name for this item type.")
documentation: DocumentationMetadataModel = Field(
description="Documentation for the module."
)
authors: AuthorsMetadataModel = Field(
description="Information about authorship for the module type."
)
context: ContextMetadataModel = Field(
description="Generic properties of this module (description, tags, labels, references, ...)."
)
def _retrieve_id(self) -> str:
return self.type_name
def _retrieve_category_id(self) -> str:
return f"type_info.{self.__class__.category_name()}"
def _retrieve_data_to_hash(self) -> Any:
return self.type_name
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
if hasattr(self, "python_class"):
table.add_row("Python class", self.python_class.create_renderable()) # type: ignore
return table
authors: AuthorsMetadataModel
pydantic-field
required
¶Information about authorship for the module type.
context: ContextMetadataModel
pydantic-field
required
¶Generic properties of this module (description, tags, labels, references, ...).
documentation: DocumentationMetadataModel
pydantic-field
required
¶Documentation for the module.
type_name: str
pydantic-field
required
¶The registered name for this item type.
category_name()
classmethod
¶Source code in kiara/models/info.py
@classmethod
@abc.abstractmethod
def category_name(cls) -> str:
pass
create_renderable(self, **config)
¶Source code in kiara/models/info.py
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
if hasattr(self, "python_class"):
table.add_row("Python class", self.python_class.create_renderable()) # type: ignore
return table
validate_doc(value)
classmethod
¶Source code in kiara/models/info.py
@validator("documentation", pre=True)
def validate_doc(cls, value):
return DocumentationMetadataModel.create(value)
KiaraTypeInfoModel (KiaraInfoModel, Generic)
pydantic-model
¶Source code in kiara/models/info.py
class KiaraTypeInfoModel(KiaraInfoModel, Generic[INFO_BASE_CLASS]):
@classmethod
@abc.abstractmethod
def create_from_type_class(
self, type_cls: Type[INFO_BASE_CLASS]
) -> "KiaraInfoModel":
pass
@classmethod
@abc.abstractmethod
def base_class(self) -> Type[INFO_BASE_CLASS]:
pass
python_class: PythonClass = Field(
description="The python class that implements this module type."
)
python_class: PythonClass
pydantic-field
required
¶The python class that implements this module type.
base_class()
classmethod
¶Source code in kiara/models/info.py
@classmethod
@abc.abstractmethod
def base_class(self) -> Type[INFO_BASE_CLASS]:
pass
create_from_type_class(type_cls)
classmethod
¶Source code in kiara/models/info.py
@classmethod
@abc.abstractmethod
def create_from_type_class(
self, type_cls: Type[INFO_BASE_CLASS]
) -> "KiaraInfoModel":
pass
TypeInfoModelGroup (InfoModelGroup, Mapping, Generic)
pydantic-model
¶Source code in kiara/models/info.py
class TypeInfoModelGroup(InfoModelGroup, Mapping[str, KiaraTypeInfoModel]):
@classmethod
@abc.abstractmethod
def base_info_class(cls) -> Type[KiaraTypeInfoModel]:
pass
@classmethod
def create_from_type_items(
cls, group_alias: Optional[str] = None, **items: Type
) -> "TypeInfoModelGroup":
type_infos = {
k: cls.base_info_class().create_from_type_class(v) for k, v in items.items()
}
data_types_info = cls.construct(group_alias=group_alias, type_infos=type_infos) # type: ignore
return data_types_info
def get_type_infos(self) -> Mapping[str, KiaraTypeInfoModel]:
return self.type_infos # type: ignore
def __getitem__(self, item: str) -> KiaraTypeInfoModel:
return self.get_type_infos()[item]
def __iter__(self):
return iter(self.get_type_infos())
def __len__(self):
return len(self.get_type_infos())
base_info_class()
classmethod
¶Source code in kiara/models/info.py
@classmethod
@abc.abstractmethod
def base_info_class(cls) -> Type[KiaraTypeInfoModel]:
pass
create_from_type_items(group_alias=None, **items)
classmethod
¶Source code in kiara/models/info.py
@classmethod
def create_from_type_items(
cls, group_alias: Optional[str] = None, **items: Type
) -> "TypeInfoModelGroup":
type_infos = {
k: cls.base_info_class().create_from_type_class(v) for k, v in items.items()
}
data_types_info = cls.construct(group_alias=group_alias, type_infos=type_infos) # type: ignore
return data_types_info
get_type_infos(self)
¶Source code in kiara/models/info.py
def get_type_infos(self) -> Mapping[str, KiaraTypeInfoModel]:
return self.type_infos # type: ignore
module
special
¶
Classes¶
KiaraModuleClass (PythonClass)
pydantic-model
¶Source code in kiara/models/module/__init__.py
class KiaraModuleClass(PythonClass):
@classmethod
def from_module(cls, module: "KiaraModule"):
item_cls = module.__class__
cls_name = item_cls.__name__
module_name = item_cls.__module__
if module_name == "builtins":
full_name = cls_name
else:
full_name = f"{item_cls.__module__}.{item_cls.__name__}"
conf: Dict[str, Any] = {
"class_name": cls_name,
"module_name": module_name,
"full_name": full_name,
}
conf["module_config"] = module.config
conf["inputs_schema"] = module.inputs_schema
conf["outputs_schema"] = module.outputs_schema
result = KiaraModuleClass.construct(**conf)
result._cls_cache = item_cls
result._module_instance_cache = module
return result
module_config: Dict[str, Any] = Field(description="The module config.")
inputs_schema: Dict[str, ValueSchema] = Field(
description="The schema for the module input(s)."
)
outputs_schema: Dict[str, ValueSchema] = Field(
description="The schema for the module output(s)."
)
_module_instance_cache: "KiaraModule" = PrivateAttr(default=None)
def get_kiara_module_instance(self) -> "KiaraModule":
if self._module_instance_cache is not None:
return self._module_instance_cache
m_cls = self.get_class()
self._module_instance_cache = m_cls(module_config=self.module_config)
return self._module_instance_cache
inputs_schema: Dict[str, kiara.models.values.value_schema.ValueSchema]
pydantic-field
required
¶The schema for the module input(s).
module_config: Dict[str, Any]
pydantic-field
required
¶The module config.
outputs_schema: Dict[str, kiara.models.values.value_schema.ValueSchema]
pydantic-field
required
¶The schema for the module output(s).
from_module(module)
classmethod
¶Source code in kiara/models/module/__init__.py
@classmethod
def from_module(cls, module: "KiaraModule"):
item_cls = module.__class__
cls_name = item_cls.__name__
module_name = item_cls.__module__
if module_name == "builtins":
full_name = cls_name
else:
full_name = f"{item_cls.__module__}.{item_cls.__name__}"
conf: Dict[str, Any] = {
"class_name": cls_name,
"module_name": module_name,
"full_name": full_name,
}
conf["module_config"] = module.config
conf["inputs_schema"] = module.inputs_schema
conf["outputs_schema"] = module.outputs_schema
result = KiaraModuleClass.construct(**conf)
result._cls_cache = item_cls
result._module_instance_cache = module
return result
get_kiara_module_instance(self)
¶Source code in kiara/models/module/__init__.py
def get_kiara_module_instance(self) -> "KiaraModule":
if self._module_instance_cache is not None:
return self._module_instance_cache
m_cls = self.get_class()
self._module_instance_cache = m_cls(module_config=self.module_config)
return self._module_instance_cache
KiaraModuleConfig (KiaraModel)
pydantic-model
¶Base class that describes the configuration a [KiaraModule][kiara.module.KiaraModule] class accepts.
This is stored in the _config_cls class attribute in each KiaraModule class.
There are two config options every KiaraModule supports:
constants, anddefaults
Constants are pre-set inputs, and users can't change them and an error is thrown if they try. Defaults are default values that override the schema defaults, and those can be overwritten by users. If both a constant and a default value is set for an input field, an error is thrown.
Source code in kiara/models/module/__init__.py
class KiaraModuleConfig(KiaraModel):
"""Base class that describes the configuration a [``KiaraModule``][kiara.module.KiaraModule] class accepts.
This is stored in the ``_config_cls`` class attribute in each ``KiaraModule`` class.
There are two config options every ``KiaraModule`` supports:
- ``constants``, and
- ``defaults``
Constants are pre-set inputs, and users can't change them and an error is thrown if they try. Defaults are default
values that override the schema defaults, and those can be overwritten by users. If both a constant and a default
value is set for an input field, an error is thrown.
"""
@classmethod
def requires_config(cls, config: Optional[Mapping[str, Any]] = None) -> bool:
"""Return whether this class can be used as-is, or requires configuration before an instance can be created."""
for field_name, field in cls.__fields__.items():
if field.required and field.default is None:
if config:
if config.get(field_name, None) is None:
return True
else:
return True
return False
_config_hash: str = PrivateAttr(default=None)
constants: Dict[str, Any] = Field(
default_factory=dict, description="Value constants for this module."
)
defaults: Dict[str, Any] = Field(
default_factory=dict, description="Value defaults for this module."
)
class Config:
extra = Extra.forbid
validate_assignment = True
def get(self, key: str) -> Any:
"""Get the value for the specified configuation key."""
if key not in self.__fields__:
raise Exception(
f"No config value '{key}' in module config class '{self.__class__.__name__}'."
)
return getattr(self, key)
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return MODULE_CONFIG_SCHEMA_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
def create_renderable(self, **config: Any) -> RenderableType:
my_table = Table(box=box.MINIMAL, show_header=False)
my_table.add_column("Field name", style="i")
my_table.add_column("Value")
for field in self.__fields__:
attr = getattr(self, field)
if isinstance(attr, str):
attr_str = attr
elif hasattr(attr, "create_renderable"):
attr_str = attr.create_renderable()
elif isinstance(attr, BaseModel):
attr_str = attr.json(option=orjson.orjson.OPT_INDENT_2)
else:
attr_str = str(attr)
my_table.add_row(field, attr_str)
return my_table
def __eq__(self, other):
if self.__class__ != other.__class__:
return False
return self.model_data_hash == other.model_data_hash
def __hash__(self):
return self.model_data_hash
constants: Dict[str, Any]
pydantic-field
¶Value constants for this module.
defaults: Dict[str, Any]
pydantic-field
¶Value defaults for this module.
Config
¶create_renderable(self, **config)
¶Source code in kiara/models/module/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
my_table = Table(box=box.MINIMAL, show_header=False)
my_table.add_column("Field name", style="i")
my_table.add_column("Value")
for field in self.__fields__:
attr = getattr(self, field)
if isinstance(attr, str):
attr_str = attr
elif hasattr(attr, "create_renderable"):
attr_str = attr.create_renderable()
elif isinstance(attr, BaseModel):
attr_str = attr.json(option=orjson.orjson.OPT_INDENT_2)
else:
attr_str = str(attr)
my_table.add_row(field, attr_str)
return my_table
get(self, key)
¶Get the value for the specified configuation key.
Source code in kiara/models/module/__init__.py
def get(self, key: str) -> Any:
"""Get the value for the specified configuation key."""
if key not in self.__fields__:
raise Exception(
f"No config value '{key}' in module config class '{self.__class__.__name__}'."
)
return getattr(self, key)
requires_config(config=None)
classmethod
¶Return whether this class can be used as-is, or requires configuration before an instance can be created.
Source code in kiara/models/module/__init__.py
@classmethod
def requires_config(cls, config: Optional[Mapping[str, Any]] = None) -> bool:
"""Return whether this class can be used as-is, or requires configuration before an instance can be created."""
for field_name, field in cls.__fields__.items():
if field.required and field.default is None:
if config:
if config.get(field_name, None) is None:
return True
else:
return True
return False
KiaraModuleConfigMetadata (KiaraModel)
pydantic-model
¶Source code in kiara/models/module/__init__.py
class KiaraModuleConfigMetadata(KiaraModel):
@classmethod
def from_config_class(
cls,
config_cls: Type[KiaraModuleConfig],
):
flat_models = get_flat_models_from_model(config_cls)
model_name_map = get_model_name_map(flat_models)
m_schema, _, _ = model_process_schema(config_cls, model_name_map=model_name_map)
fields = m_schema["properties"]
config_values = {}
for field_name, details in fields.items():
type_str = "-- n/a --"
if "type" in details.keys():
type_str = details["type"]
desc = details.get("description", DEFAULT_NO_DESC_VALUE)
default = config_cls.__fields__[field_name].default
if default is None:
if callable(config_cls.__fields__[field_name].default_factory):
default = config_cls.__fields__[field_name].default_factory() # type: ignore
req = config_cls.__fields__[field_name].required
config_values[field_name] = ValueTypeAndDescription(
description=desc, type=type_str, value_default=default, required=req
)
python_cls = PythonClass.from_class(config_cls)
return KiaraModuleConfigMetadata(
python_class=python_cls, config_values=config_values
)
python_class: PythonClass = Field(description="The config model python class.")
config_values: Dict[str, ValueTypeAndDescription] = Field(
description="The available configuration values."
)
def _retrieve_id(self) -> str:
return self.python_class.model_id
def _retrieve_category_id(self) -> str:
return MODULE_CONFIG_METADATA_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.python_class.model_id
config_values: Dict[str, kiara.models.module.ValueTypeAndDescription]
pydantic-field
required
¶The available configuration values.
python_class: PythonClass
pydantic-field
required
¶The config model python class.
from_config_class(config_cls)
classmethod
¶Source code in kiara/models/module/__init__.py
@classmethod
def from_config_class(
cls,
config_cls: Type[KiaraModuleConfig],
):
flat_models = get_flat_models_from_model(config_cls)
model_name_map = get_model_name_map(flat_models)
m_schema, _, _ = model_process_schema(config_cls, model_name_map=model_name_map)
fields = m_schema["properties"]
config_values = {}
for field_name, details in fields.items():
type_str = "-- n/a --"
if "type" in details.keys():
type_str = details["type"]
desc = details.get("description", DEFAULT_NO_DESC_VALUE)
default = config_cls.__fields__[field_name].default
if default is None:
if callable(config_cls.__fields__[field_name].default_factory):
default = config_cls.__fields__[field_name].default_factory() # type: ignore
req = config_cls.__fields__[field_name].required
config_values[field_name] = ValueTypeAndDescription(
description=desc, type=type_str, value_default=default, required=req
)
python_cls = PythonClass.from_class(config_cls)
return KiaraModuleConfigMetadata(
python_class=python_cls, config_values=config_values
)
KiaraModuleTypeInfo (KiaraTypeInfoModel)
pydantic-model
¶Source code in kiara/models/module/__init__.py
class KiaraModuleTypeInfo(KiaraTypeInfoModel["KiaraModule"]):
@classmethod
def create_from_type_class(
cls, type_cls: Type["KiaraModule"]
) -> "KiaraModuleTypeInfo":
module_attrs = cls.extract_module_attributes(module_cls=type_cls)
return cls.construct(**module_attrs)
@classmethod
def base_class(self) -> Type["KiaraModule"]:
from kiara.modules import KiaraModule
return KiaraModule
@classmethod
def category_name(cls) -> str:
return "module"
@classmethod
def extract_module_attributes(
self, module_cls: Type["KiaraModule"]
) -> Dict[str, Any]:
if not hasattr(module_cls, "process"):
raise Exception(f"Module class '{module_cls}' misses 'process' method.")
proc_src = textwrap.dedent(inspect.getsource(module_cls.process)) # type: ignore
authors_md = AuthorsMetadataModel.from_class(module_cls)
doc = DocumentationMetadataModel.from_class_doc(module_cls)
python_class = PythonClass.from_class(module_cls)
properties_md = ContextMetadataModel.from_class(module_cls)
config = KiaraModuleConfigMetadata.from_config_class(module_cls._config_cls)
return {
"type_name": module_cls._module_type_name, # type: ignore
"documentation": doc,
"authors": authors_md,
"context": properties_md,
"python_class": python_class,
"config": config,
"process_src": proc_src,
}
process_src: str = Field(
description="The source code of the process method of the module."
)
def create_renderable(self, **config: Any) -> RenderableType:
include_config_schema = config.get("include_config_schema", True)
include_src = config.get("include_src", True)
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
if include_config_schema:
config_cls = self.python_class.get_class()._config_cls # type: ignore
from kiara.utils.output import create_table_from_base_model_cls
table.add_row(
"Module config schema", create_table_from_base_model_cls(config_cls)
)
table.add_row("Python class", self.python_class.create_renderable())
if include_src:
_config = Syntax(self.process_src, "python", background_color="default")
table.add_row("Processing source code", Panel(_config, box=box.HORIZONTALS))
return table
process_src: str
pydantic-field
required
¶The source code of the process method of the module.
base_class()
classmethod
¶Source code in kiara/models/module/__init__.py
@classmethod
def base_class(self) -> Type["KiaraModule"]:
from kiara.modules import KiaraModule
return KiaraModule
category_name()
classmethod
¶Source code in kiara/models/module/__init__.py
@classmethod
def category_name(cls) -> str:
return "module"
create_from_type_class(type_cls)
classmethod
¶Source code in kiara/models/module/__init__.py
@classmethod
def create_from_type_class(
cls, type_cls: Type["KiaraModule"]
) -> "KiaraModuleTypeInfo":
module_attrs = cls.extract_module_attributes(module_cls=type_cls)
return cls.construct(**module_attrs)
create_renderable(self, **config)
¶Source code in kiara/models/module/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
include_config_schema = config.get("include_config_schema", True)
include_src = config.get("include_src", True)
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
if include_config_schema:
config_cls = self.python_class.get_class()._config_cls # type: ignore
from kiara.utils.output import create_table_from_base_model_cls
table.add_row(
"Module config schema", create_table_from_base_model_cls(config_cls)
)
table.add_row("Python class", self.python_class.create_renderable())
if include_src:
_config = Syntax(self.process_src, "python", background_color="default")
table.add_row("Processing source code", Panel(_config, box=box.HORIZONTALS))
return table
extract_module_attributes(module_cls)
classmethod
¶Source code in kiara/models/module/__init__.py
@classmethod
def extract_module_attributes(
self, module_cls: Type["KiaraModule"]
) -> Dict[str, Any]:
if not hasattr(module_cls, "process"):
raise Exception(f"Module class '{module_cls}' misses 'process' method.")
proc_src = textwrap.dedent(inspect.getsource(module_cls.process)) # type: ignore
authors_md = AuthorsMetadataModel.from_class(module_cls)
doc = DocumentationMetadataModel.from_class_doc(module_cls)
python_class = PythonClass.from_class(module_cls)
properties_md = ContextMetadataModel.from_class(module_cls)
config = KiaraModuleConfigMetadata.from_config_class(module_cls._config_cls)
return {
"type_name": module_cls._module_type_name, # type: ignore
"documentation": doc,
"authors": authors_md,
"context": properties_md,
"python_class": python_class,
"config": config,
"process_src": proc_src,
}
ModuleTypeClassesInfo (TypeInfoModelGroup)
pydantic-model
¶Source code in kiara/models/module/__init__.py
class ModuleTypeClassesInfo(TypeInfoModelGroup):
@classmethod
def base_info_class(cls) -> Type[KiaraTypeInfoModel]:
return KiaraModuleTypeInfo
type_name: Literal["module_type"] = "module_type"
type_infos: Mapping[str, KiaraModuleTypeInfo] = Field(
description="The module type info instances for each type."
)
type_infos: Mapping[str, kiara.models.module.KiaraModuleTypeInfo]
pydantic-field
required
¶The module type info instances for each type.
type_name: Literal['module_type']
pydantic-field
¶base_info_class()
classmethod
¶Source code in kiara/models/module/__init__.py
@classmethod
def base_info_class(cls) -> Type[KiaraTypeInfoModel]:
return KiaraModuleTypeInfo
ValueTypeAndDescription (BaseModel)
pydantic-model
¶Source code in kiara/models/module/__init__.py
class ValueTypeAndDescription(BaseModel):
description: str = Field(description="The description for the value.")
type: str = Field(description="The value type.")
value_default: Any = Field(description="Default for the value.", default=None)
required: bool = Field(description="Whether this value is required")
calculate_class_doc_url(base_url, module_type_name)
¶Source code in kiara/models/module/__init__.py
def calculate_class_doc_url(base_url: str, module_type_name: str):
if base_url.endswith("/"):
base_url = base_url[0:-1]
module_type_name = module_type_name.replace(".", "")
url = f"{base_url}/latest/modules_list/#{module_type_name}"
return url
calculate_class_source_url(base_url, python_class_info, branch='main')
¶Source code in kiara/models/module/__init__.py
def calculate_class_source_url(
base_url: str, python_class_info: PythonClass, branch: str = "main"
):
if base_url.endswith("/"):
base_url = base_url[0:-1]
m = python_class_info.get_python_module()
m_file = m.__file__
assert m_file is not None
base_url = f"{base_url}/blob/{branch}/src/{python_class_info.module_name.replace('.', '/')}"
if m_file.endswith("__init__.py"):
url = f"{base_url}/__init__.py"
else:
url = f"{base_url}.py"
return url
Modules¶
destiniy
¶
Destiny (Manifest)
pydantic-model
¶A destiny is basically a link to a potential future transformation result involving one or several values as input.
It is immutable, once executed, each of the input values can only have one destiny with a specific alias. This is similar to what is usually called a 'future' in programming languages, but more deterministic, sorta.
Source code in kiara/models/module/destiniy.py
class Destiny(Manifest):
"""A destiny is basically a link to a potential future transformation result involving one or several values as input.
It is immutable, once executed, each of the input values can only have one destiny with a specific alias.
This is similar to what is usually called a 'future' in programming languages, but more deterministic, sorta.
"""
@classmethod
def create_from_values(
cls,
kiara: "Kiara",
destiny_alias: str,
values: Mapping[str, uuid.UUID],
manifest: Manifest,
result_field_name: Optional[str] = None,
):
module = kiara.create_module(manifest=manifest)
if result_field_name is None:
if len(module.outputs_schema) != 1:
raise Exception(
f"Can't determine result field name for module, not provided, and multiple outputs available for module '{module.module_type_name}': {', '.join(module.outputs_schema.keys())}."
)
result_field_name = next(iter(module.outputs_schema.keys()))
result_schema = module.outputs_schema.get(result_field_name, None)
if result_schema is None:
raise Exception(
f"Can't determine result schema for module '{module.module_type_name}', result field '{result_field_name}' not available. Available field: {', '.join(module.outputs_schema.keys())}"
)
fixed_inputs = {}
deferred_inputs: Dict[str, None] = {}
for field in module.inputs_schema.keys():
if field in values.keys():
fixed_inputs[field] = values[field]
else:
deferred_inputs[field] = None
module_details = KiaraModuleClass.from_module(module=module)
# TODO: check whether it'd be better to 'resolve' the module config, as this might change the resulting hash
destiny_id: uuid.UUID = ID_REGISTRY.generate(obj_type=Destiny)
destiny = Destiny(
destiny_id=destiny_id,
destiny_alias=destiny_alias,
module_details=module_details,
module_type=manifest.module_type,
module_config=manifest.module_config,
result_field_name=result_field_name,
result_schema=result_schema,
fixed_inputs=fixed_inputs,
inputs_schema=dict(module.inputs_schema),
deferred_inputs=deferred_inputs,
)
destiny._module = module
ID_REGISTRY.update_metadata(destiny_id, obj=destiny)
return destiny
destiny_id: uuid.UUID = Field(description="The id of this destiny.")
destiny_alias: str = Field(description="The path to (the) destiny.")
module_details: KiaraModuleClass = Field(
description="The class of the underlying module."
)
fixed_inputs: Dict[str, uuid.UUID] = Field(
description="Inputs that are known in advance."
)
inputs_schema: Dict[str, ValueSchema] = Field(
description="The schemas of all deferred input fields."
)
deferred_inputs: Dict[str, Optional[uuid.UUID]] = Field(
description="Potentially required external inputs that are needed for this destiny to materialize."
)
result_field_name: str = Field(description="The name of the result field.")
result_schema: ValueSchema = Field(description="The value schema of the result.")
result_value_id: Optional[uuid.UUID] = Field(
description="The value id of the result."
)
_is_stored: bool = PrivateAttr(default=False)
_job_id: Optional[uuid.UUID] = PrivateAttr(default=None)
_merged_inputs: Optional[Dict[str, uuid.UUID]] = PrivateAttr(default=None)
_job_config_hash: Optional[int] = PrivateAttr(default=None)
_module: Optional["KiaraModule"] = PrivateAttr(default=None)
def _retrieve_id(self) -> str:
return str(self.destiny_id)
def _retrieve_category_id(self) -> str:
return DESTINY_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.destiny_id
@property
def job_config_hash(self) -> int:
if self._job_config_hash is None:
self._job_config_hash = self._retrieve_job_config_hash()
return self._job_config_hash
@property
def merged_inputs(self) -> Mapping[str, uuid.UUID]:
if self._merged_inputs is not None:
return self._merged_inputs
result = copy.copy(self.fixed_inputs)
missing = []
for k in self.inputs_schema.keys():
if k in self.fixed_inputs.keys():
if k in self.deferred_inputs.keys():
raise Exception(
f"Destiny input field '{k}' present in both fixed and deferred inputs, this is invalid."
)
else:
continue
v = self.deferred_inputs.get(k, None)
if v is None or isinstance(v, SpecialValue):
missing.append(k)
else:
result[k] = v
if missing:
raise Exception(
f"Destiny not valid (yet), missing inputs: {', '.join(missing)}"
)
self._merged_inputs = result
return self._merged_inputs
@property
def module(self) -> "KiaraModule":
if self._module is None:
m_cls = self.module_details.get_class()
self._module = m_cls(module_config=self.module_config)
return self._module
def _retrieve_job_config_hash(self) -> int:
obj = {"module_config": self.manifest_data, "inputs": self.merged_inputs}
return compute_hash(obj)
deferred_inputs: Dict[str, Optional[uuid.UUID]]
pydantic-field
required
¶Potentially required external inputs that are needed for this destiny to materialize.
destiny_alias: str
pydantic-field
required
¶The path to (the) destiny.
destiny_id: UUID
pydantic-field
required
¶The id of this destiny.
fixed_inputs: Dict[str, uuid.UUID]
pydantic-field
required
¶Inputs that are known in advance.
inputs_schema: Dict[str, kiara.models.values.value_schema.ValueSchema]
pydantic-field
required
¶The schemas of all deferred input fields.
job_config_hash: int
property
readonly
¶merged_inputs: Mapping[str, uuid.UUID]
property
readonly
¶module: KiaraModule
property
readonly
¶module_details: KiaraModuleClass
pydantic-field
required
¶The class of the underlying module.
result_field_name: str
pydantic-field
required
¶The name of the result field.
result_schema: ValueSchema
pydantic-field
required
¶The value schema of the result.
result_value_id: UUID
pydantic-field
¶The value id of the result.
create_from_values(kiara, destiny_alias, values, manifest, result_field_name=None)
classmethod
¶Source code in kiara/models/module/destiniy.py
@classmethod
def create_from_values(
cls,
kiara: "Kiara",
destiny_alias: str,
values: Mapping[str, uuid.UUID],
manifest: Manifest,
result_field_name: Optional[str] = None,
):
module = kiara.create_module(manifest=manifest)
if result_field_name is None:
if len(module.outputs_schema) != 1:
raise Exception(
f"Can't determine result field name for module, not provided, and multiple outputs available for module '{module.module_type_name}': {', '.join(module.outputs_schema.keys())}."
)
result_field_name = next(iter(module.outputs_schema.keys()))
result_schema = module.outputs_schema.get(result_field_name, None)
if result_schema is None:
raise Exception(
f"Can't determine result schema for module '{module.module_type_name}', result field '{result_field_name}' not available. Available field: {', '.join(module.outputs_schema.keys())}"
)
fixed_inputs = {}
deferred_inputs: Dict[str, None] = {}
for field in module.inputs_schema.keys():
if field in values.keys():
fixed_inputs[field] = values[field]
else:
deferred_inputs[field] = None
module_details = KiaraModuleClass.from_module(module=module)
# TODO: check whether it'd be better to 'resolve' the module config, as this might change the resulting hash
destiny_id: uuid.UUID = ID_REGISTRY.generate(obj_type=Destiny)
destiny = Destiny(
destiny_id=destiny_id,
destiny_alias=destiny_alias,
module_details=module_details,
module_type=manifest.module_type,
module_config=manifest.module_config,
result_field_name=result_field_name,
result_schema=result_schema,
fixed_inputs=fixed_inputs,
inputs_schema=dict(module.inputs_schema),
deferred_inputs=deferred_inputs,
)
destiny._module = module
ID_REGISTRY.update_metadata(destiny_id, obj=destiny)
return destiny
jobs
¶
ActiveJob (KiaraModel)
pydantic-model
¶Source code in kiara/models/module/jobs.py
class ActiveJob(KiaraModel):
job_id: uuid.UUID = Field(description="The job id.")
job_config: JobConfig = Field(description="The job details.")
status: JobStatus = Field(
description="The current status of the job.", default=JobStatus.CREATED
)
job_log: JobLog = Field(description="The lob jog.")
submitted: datetime = Field(
description="When the job was submitted.", default_factory=datetime.now
)
started: Optional[datetime] = Field(
description="When the job was started.", default=None
)
finished: Optional[datetime] = Field(
description="When the job was finished.", default=None
)
results: Optional[Dict[str, uuid.UUID]] = Field(description="The result(s).")
error: Optional[str] = Field(description="Potential error message.")
_exception: Optional[Exception] = PrivateAttr(default=None)
def _retrieve_id(self) -> str:
return str(self.job_id)
def _retrieve_category_id(self) -> str:
return JOB_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.job_id
@property
def exception(self) -> Optional[Exception]:
return self._exception
@property
def runtime(self) -> Optional[float]:
if self.started is None or self.finished is None:
return None
runtime = self.finished - self.started
return runtime.total_seconds()
error: str
pydantic-field
¶Potential error message.
exception: Optional[Exception]
property
readonly
¶finished: datetime
pydantic-field
¶When the job was finished.
job_config: JobConfig
pydantic-field
required
¶The job details.
job_id: UUID
pydantic-field
required
¶The job id.
job_log: JobLog
pydantic-field
required
¶The lob jog.
results: Dict[str, uuid.UUID]
pydantic-field
¶The result(s).
runtime: Optional[float]
property
readonly
¶started: datetime
pydantic-field
¶When the job was started.
status: JobStatus
pydantic-field
¶The current status of the job.
submitted: datetime
pydantic-field
¶When the job was submitted.
JobConfig (InputsManifest)
pydantic-model
¶Source code in kiara/models/module/jobs.py
class JobConfig(InputsManifest):
@classmethod
def create_from_module(
cls,
data_registry: "DataRegistry",
module: "KiaraModule",
inputs: Mapping[str, Any],
):
augmented = module.augment_module_inputs(inputs=inputs)
values = data_registry.create_valueset(
data=augmented, schema=module.inputs_schema
)
invalid = values.check_invalid()
if invalid:
raise InvalidValuesException(invalid_values=invalid)
value_ids = values.get_all_value_ids()
return JobConfig.construct(
module_type=module.module_type_name,
module_config=module.config.dict(),
inputs=value_ids,
)
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return JOB_CONFIG_TYPE_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return {"manifest": self.manifest_data, "inputs": self.inputs_hash}
create_from_module(data_registry, module, inputs)
classmethod
¶Source code in kiara/models/module/jobs.py
@classmethod
def create_from_module(
cls,
data_registry: "DataRegistry",
module: "KiaraModule",
inputs: Mapping[str, Any],
):
augmented = module.augment_module_inputs(inputs=inputs)
values = data_registry.create_valueset(
data=augmented, schema=module.inputs_schema
)
invalid = values.check_invalid()
if invalid:
raise InvalidValuesException(invalid_values=invalid)
value_ids = values.get_all_value_ids()
return JobConfig.construct(
module_type=module.module_type_name,
module_config=module.config.dict(),
inputs=value_ids,
)
JobLog (BaseModel)
pydantic-model
¶Source code in kiara/models/module/jobs.py
class JobLog(BaseModel):
log: List[LogMessage] = Field(
description="The logs for this job.", default_factory=list
)
percent_finished: int = Field(
description="Describes how much of the job is finished. A negative number means the module does not support progress tracking.",
default=-1,
)
def add_log(self, msg: str, log_level: int = logging.DEBUG):
_msg = LogMessage(msg=msg, log_level=log_level)
self.log.append(_msg)
log: List[kiara.models.module.jobs.LogMessage]
pydantic-field
¶The logs for this job.
percent_finished: int
pydantic-field
¶Describes how much of the job is finished. A negative number means the module does not support progress tracking.
add_log(self, msg, log_level=10)
¶Source code in kiara/models/module/jobs.py
def add_log(self, msg: str, log_level: int = logging.DEBUG):
_msg = LogMessage(msg=msg, log_level=log_level)
self.log.append(_msg)
JobRecord (JobConfig)
pydantic-model
¶Source code in kiara/models/module/jobs.py
class JobRecord(JobConfig):
@classmethod
def from_active_job(self, active_job: ActiveJob):
assert active_job.status == JobStatus.SUCCESS
assert active_job.results is not None
job_details = JobRuntimeDetails.construct(
job_log=active_job.job_log,
submitted=active_job.submitted,
started=active_job.started, # type: ignore
finished=active_job.finished, # type: ignore
runtime=active_job.runtime, # type: ignore
)
job_record = JobRecord.construct(
job_id=active_job.job_id,
module_type=active_job.job_config.module_type,
module_config=active_job.job_config.module_config,
inputs=active_job.job_config.inputs,
outputs=active_job.results,
runtime_details=job_details,
)
return job_record
job_id: uuid.UUID = Field(description="The globally unique id for this job.")
outputs: Dict[str, uuid.UUID] = Field(description="References to the job outputs.")
runtime_details: Optional[JobRuntimeDetails] = Field(
description="Runtime details for the job."
)
_is_stored: bool = PrivateAttr(default=None)
_outputs_hash: Optional[int] = PrivateAttr(default=None)
def _retrieve_category_id(self) -> str:
return JOB_RECORD_TYPE_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return {
"manifest_hash": self.manifest_hash,
"inputs": self.inputs,
"outputs": self.outputs,
}
@property
def outputs_hash(self) -> int:
if self._outputs_hash is not None:
return self._outputs_hash
obj = self.outputs
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
self._outputs_hash = h[obj]
return self._outputs_hash
job_id: UUID
pydantic-field
required
¶The globally unique id for this job.
outputs: Dict[str, uuid.UUID]
pydantic-field
required
¶References to the job outputs.
outputs_hash: int
property
readonly
¶runtime_details: JobRuntimeDetails
pydantic-field
¶Runtime details for the job.
from_active_job(active_job)
classmethod
¶Source code in kiara/models/module/jobs.py
@classmethod
def from_active_job(self, active_job: ActiveJob):
assert active_job.status == JobStatus.SUCCESS
assert active_job.results is not None
job_details = JobRuntimeDetails.construct(
job_log=active_job.job_log,
submitted=active_job.submitted,
started=active_job.started, # type: ignore
finished=active_job.finished, # type: ignore
runtime=active_job.runtime, # type: ignore
)
job_record = JobRecord.construct(
job_id=active_job.job_id,
module_type=active_job.job_config.module_type,
module_config=active_job.job_config.module_config,
inputs=active_job.job_config.inputs,
outputs=active_job.results,
runtime_details=job_details,
)
return job_record
JobRuntimeDetails (BaseModel)
pydantic-model
¶Source code in kiara/models/module/jobs.py
class JobRuntimeDetails(BaseModel):
# @classmethod
# def from_manifest(
# cls,
# manifest: Manifest,
# inputs: Mapping[str, Value],
# outputs: Mapping[str, Value],
# ):
#
# return JobRecord(
# module_type=manifest.module_type,
# module_config=manifest.module_config,
# inputs={k: v.value_id for k, v in inputs.items()},
# outputs={k: v.value_id for k, v in outputs.items()},
# )
job_log: JobLog = Field(description="The lob jog.")
submitted: datetime = Field(description="When the job was submitted.")
started: datetime = Field(description="When the job was started.")
finished: datetime = Field(description="When the job was finished.")
runtime: float = Field(description="The duration of the job.")
finished: datetime
pydantic-field
required
¶When the job was finished.
job_log: JobLog
pydantic-field
required
¶The lob jog.
runtime: float
pydantic-field
required
¶The duration of the job.
started: datetime
pydantic-field
required
¶When the job was started.
submitted: datetime
pydantic-field
required
¶When the job was submitted.
JobStatus (Enum)
¶
LogMessage (BaseModel)
pydantic-model
¶Source code in kiara/models/module/jobs.py
class LogMessage(BaseModel):
timestamp: datetime = Field(
description="The time the message was logged.", default_factory=datetime.now
)
log_level: int = Field(description="The log level.")
msg: str = Field(description="The log message")
manifest
¶
InputsManifest (Manifest)
pydantic-model
¶Source code in kiara/models/module/manifest.py
class InputsManifest(Manifest):
inputs: Mapping[str, uuid.UUID] = Field(
description="A map of all the input fields and value references."
)
_inputs_hash: Optional[int] = PrivateAttr(default=None)
_jobs_hash: Optional[int] = PrivateAttr(default=None)
@property
def job_hash(self) -> int:
if self._jobs_hash is not None:
return self._jobs_hash
obj = {"manifest": self.manifest_hash, "inputs": self.inputs_hash}
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
self._jobs_hash = h[obj]
return self._jobs_hash
@validator("inputs")
def replace_none_values(cls, value):
result = {}
for k, v in value.items():
if v is None:
v = NONE_VALUE_ID
result[k] = v
return result
@property
def inputs_hash(self) -> int:
if self._inputs_hash is not None:
return self._inputs_hash
if self.module_type == NO_MODULE_TYPE and not self.inputs:
self._inputs_hash = 0
else:
h = DeepHash(self.inputs, hasher=KIARA_HASH_FUNCTION)
self._inputs_hash = h[self.inputs]
return self._inputs_hash
inputs: Mapping[str, uuid.UUID]
pydantic-field
required
¶A map of all the input fields and value references.
inputs_hash: int
property
readonly
¶job_hash: int
property
readonly
¶replace_none_values(value)
classmethod
¶Source code in kiara/models/module/manifest.py
@validator("inputs")
def replace_none_values(cls, value):
result = {}
for k, v in value.items():
if v is None:
v = NONE_VALUE_ID
result[k] = v
return result
Manifest (KiaraModel)
pydantic-model
¶A class to hold the type and configuration for a module instance.
Source code in kiara/models/module/manifest.py
class Manifest(KiaraModel):
"""A class to hold the type and configuration for a module instance."""
class Config:
extra = Extra.forbid
validate_all = True
_manifest_data: Optional[Mapping[str, Any]] = PrivateAttr(default=None)
_manifest_hash: Optional[int] = PrivateAttr(default=None)
module_type: str = Field(description="The module type.")
module_config: Mapping[str, Any] = Field(
default_factory=dict, description="The configuration for the module."
)
# python_class: PythonClass = Field(description="The python class that implements this module.")
# doc: DocumentationMetadataModel = Field(
# description="Documentation for this module instance.", default=None
# )
@property
def manifest_data(self):
"""The configuration data for this module instance."""
if self._manifest_data is not None:
return self._manifest_data
self._manifest_data = {
"module_type": self.module_type,
"module_config": self.module_config,
}
return self._manifest_data
def manifest_data_as_json(self):
return self.json(include={"module_type", "module_config"})
@property
def manifest_hash(self) -> int:
"""The hash for the inherent module config (composted of type and render_config data).
Not that this can (but might not) be different to the `model_data_hash`.
"""
if self._manifest_hash is not None:
return self._manifest_hash
h = DeepHash(self.manifest_data, hasher=KIARA_HASH_FUNCTION)
self._manifest_hash = h[self.manifest_data]
return self._manifest_hash
def _retrieve_data_to_hash(self) -> Any:
return self.manifest_data
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return MODULE_CONFIG_CATEGORY_ID
def create_renderable(self, **config: Any) -> RenderableType:
"""Create a renderable for this module configuration."""
data = self.dict(exclude_none=True)
conf = Syntax(
orjson_dumps(data, option=orjson.OPT_INDENT_2),
"json",
background_color="default",
)
return conf
def __repr__(self):
return f"{self.__class__.__name__}(module_type={self.module_type}, module_config={self.module_config})"
def __str__(self):
return self.__repr__()
manifest_data
property
readonly
¶The configuration data for this module instance.
manifest_hash: int
property
readonly
¶The hash for the inherent module config (composted of type and render_config data).
Not that this can (but might not) be different to the model_data_hash.
module_config: Mapping[str, Any]
pydantic-field
¶The configuration for the module.
module_type: str
pydantic-field
required
¶The module type.
Config
¶create_renderable(self, **config)
¶Create a renderable for this module configuration.
Source code in kiara/models/module/manifest.py
def create_renderable(self, **config: Any) -> RenderableType:
"""Create a renderable for this module configuration."""
data = self.dict(exclude_none=True)
conf = Syntax(
orjson_dumps(data, option=orjson.OPT_INDENT_2),
"json",
background_color="default",
)
return conf
manifest_data_as_json(self)
¶Source code in kiara/models/module/manifest.py
def manifest_data_as_json(self):
return self.json(include={"module_type", "module_config"})
operation
¶logger
¶
BaseOperationDetails (OperationDetails)
pydantic-model
¶Source code in kiara/models/module/operation.py
class BaseOperationDetails(OperationDetails):
_op_schema: OperationSchema = PrivateAttr(default=None)
def retrieve_inputs_schema(cls) -> ValueSetSchema:
raise NotImplementedError()
def retrieve_outputs_schema(cls) -> ValueSetSchema:
raise NotImplementedError()
def get_operation_schema(self) -> OperationSchema:
if self._op_schema is not None:
return self._op_schema
self._op_schema = OperationSchema(
alias=self.__class__.__name__,
inputs_schema=self.retrieve_inputs_schema(),
outputs_schema=self.retrieve_outputs_schema(),
)
return self._op_schema
get_operation_schema(self)
¶Source code in kiara/models/module/operation.py
def get_operation_schema(self) -> OperationSchema:
if self._op_schema is not None:
return self._op_schema
self._op_schema = OperationSchema(
alias=self.__class__.__name__,
inputs_schema=self.retrieve_inputs_schema(),
outputs_schema=self.retrieve_outputs_schema(),
)
return self._op_schema
retrieve_inputs_schema(cls)
¶Source code in kiara/models/module/operation.py
def retrieve_inputs_schema(cls) -> ValueSetSchema:
raise NotImplementedError()
retrieve_outputs_schema(cls)
¶Source code in kiara/models/module/operation.py
def retrieve_outputs_schema(cls) -> ValueSetSchema:
raise NotImplementedError()
ManifestOperationConfig (OperationConfig)
pydantic-model
¶Source code in kiara/models/module/operation.py
class ManifestOperationConfig(OperationConfig):
module_type: str = Field(description="The module type.")
module_config: Dict[str, Any] = Field(
default_factory=dict, description="The configuration for the module."
)
def retrieve_module_type(self, kiara: "Kiara") -> str:
return self.module_type
def retrieve_module_config(self, kiara: "Kiara") -> Mapping[str, Any]:
return self.module_config
module_config: Dict[str, Any]
pydantic-field
¶The configuration for the module.
module_type: str
pydantic-field
required
¶The module type.
retrieve_module_config(self, kiara)
¶Source code in kiara/models/module/operation.py
def retrieve_module_config(self, kiara: "Kiara") -> Mapping[str, Any]:
return self.module_config
retrieve_module_type(self, kiara)
¶Source code in kiara/models/module/operation.py
def retrieve_module_type(self, kiara: "Kiara") -> str:
return self.module_type
Operation (Manifest)
pydantic-model
¶Source code in kiara/models/module/operation.py
class Operation(Manifest):
@classmethod
def create_from_module(cls, module: KiaraModule) -> "Operation":
from kiara.operations.included_core_operations import (
CustomModuleOperationDetails,
)
op_id = f"{module.module_type_name}._{module.module_instance_hash}"
details = CustomModuleOperationDetails.create_from_module(module=module)
operation = Operation(
module_type=module.module_type_name,
module_config=module.config.dict(),
operation_id=op_id,
operation_details=details,
module_details=KiaraModuleClass.from_module(module),
doc=DocumentationMetadataModel.from_class_doc(module.__class__),
)
operation._module = module
return operation
operation_id: str = Field(description="The (unique) id of this operation.")
operation_details: OperationDetails = Field(
description="The operation specific details of this operation."
)
doc: DocumentationMetadataModel = Field(
description="Documentation for this operation."
)
module_details: KiaraModuleClass = Field(
description="The class of the underlying module."
)
_module: Optional["KiaraModule"] = PrivateAttr(default=None)
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return OPERATION_CATEOGORY_ID
def _retrieve_data_to_hash(self) -> Any:
raise NotImplementedError()
@property
def module(self) -> "KiaraModule":
if self._module is None:
m_cls = self.module_details.get_class()
self._module = m_cls(module_config=self.module_config)
return self._module
@property
def inputs_schema(self) -> Mapping[str, ValueSchema]:
return self.operation_details.inputs_schema
@property
def outputs_schema(self) -> Mapping[str, ValueSchema]:
return self.operation_details.outputs_schema
def prepare_job_config(
self, kiara: "Kiara", inputs: Mapping[str, Any]
) -> JobConfig:
augmented_inputs = (
self.operation_details.get_operation_schema().augment_module_inputs(
inputs=inputs
)
)
module_inputs = self.operation_details.create_module_inputs(
inputs=augmented_inputs
)
job_config = kiara.job_registry.prepare_job_config(
manifest=self, inputs=module_inputs
)
return job_config
def run(self, kiara: "Kiara", inputs: Any) -> ValueMap:
logger.debug("run.operation", operation_id=self.operation_id)
job_config = self.prepare_job_config(kiara=kiara, inputs=inputs)
job_id = kiara.job_registry.execute_job(job_config=job_config)
outputs: ValueMap = kiara.job_registry.retrieve_result(job_id=job_id)
result = self.process_job_outputs(outputs=outputs)
return result
def process_job_outputs(self, outputs: ValueMap) -> ValueMap:
op_outputs = self.operation_details.create_operation_outputs(outputs=outputs)
value_set = ValueMapReadOnly(value_items=op_outputs, values_schema=self.outputs_schema) # type: ignore
return value_set
# def run(self, _attach_lineage: bool = True, **inputs: Any) -> ValueMap:
#
# return self.module.run(_attach_lineage=_attach_lineage, **inputs)
def create_renderable(self, **config: Any) -> RenderableType:
"""Create a printable overview of this operations details.
Available render_config options:
- 'include_full_doc' (default: True): whether to include the full documentation, or just a description
- 'include_src' (default: False): whether to include the module source code
"""
include_full_doc = config.get("include_full_doc", True)
include_src = config.get("include_src", False)
include_inputs = config.get("include_inputs", True)
include_outputs = config.get("include_outputs", True)
include_module_details = config.get("include_moduel_details", True)
table = Table(box=box.SIMPLE, show_header=False, show_lines=True)
table.add_column("Property", style="i")
table.add_column("Value")
if self.doc:
if include_full_doc:
table.add_row("Documentation", self.doc.full_doc)
else:
table.add_row("Description", self.doc.description)
# module_type_md = self.module.get_type_metadata()
if include_inputs:
inputs_table = create_table_from_field_schemas(
_add_required=True,
_add_default=True,
_show_header=True,
_constants=None,
**self.operation_details.inputs_schema,
)
table.add_row("Inputs", inputs_table)
if include_outputs:
outputs_table = create_table_from_field_schemas(
_add_required=False,
_add_default=False,
_show_header=True,
_constants=None,
**self.operation_details.outputs_schema,
)
table.add_row("Outputs", outputs_table)
if include_module_details:
table.add_row("Module type", self.module_type)
module_config = self.module.config.json(option=orjson.OPT_INDENT_2)
conf = Syntax(
module_config,
"json",
background_color="default",
)
table.add_row("Module config", conf)
module_type_md = KiaraModuleTypeInfo.create_from_type_class(
self.module_details.get_class() # type: ignore
)
desc = module_type_md.documentation.description
module_md = module_type_md.create_renderable(
include_doc=False, include_src=False, include_config_schema=False
)
m_md = Group(desc, module_md)
table.add_row("Module metadata", m_md)
if include_src:
table.add_row("Source code", module_type_md.process_src)
return table
doc: DocumentationMetadataModel
pydantic-field
required
¶Documentation for this operation.
inputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
property
readonly
¶module: KiaraModule
property
readonly
¶module_details: KiaraModuleClass
pydantic-field
required
¶The class of the underlying module.
operation_details: OperationDetails
pydantic-field
required
¶The operation specific details of this operation.
operation_id: str
pydantic-field
required
¶The (unique) id of this operation.
outputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
property
readonly
¶create_from_module(module)
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def create_from_module(cls, module: KiaraModule) -> "Operation":
from kiara.operations.included_core_operations import (
CustomModuleOperationDetails,
)
op_id = f"{module.module_type_name}._{module.module_instance_hash}"
details = CustomModuleOperationDetails.create_from_module(module=module)
operation = Operation(
module_type=module.module_type_name,
module_config=module.config.dict(),
operation_id=op_id,
operation_details=details,
module_details=KiaraModuleClass.from_module(module),
doc=DocumentationMetadataModel.from_class_doc(module.__class__),
)
operation._module = module
return operation
create_renderable(self, **config)
¶Create a printable overview of this operations details.
Available render_config options: - 'include_full_doc' (default: True): whether to include the full documentation, or just a description - 'include_src' (default: False): whether to include the module source code
Source code in kiara/models/module/operation.py
def create_renderable(self, **config: Any) -> RenderableType:
"""Create a printable overview of this operations details.
Available render_config options:
- 'include_full_doc' (default: True): whether to include the full documentation, or just a description
- 'include_src' (default: False): whether to include the module source code
"""
include_full_doc = config.get("include_full_doc", True)
include_src = config.get("include_src", False)
include_inputs = config.get("include_inputs", True)
include_outputs = config.get("include_outputs", True)
include_module_details = config.get("include_moduel_details", True)
table = Table(box=box.SIMPLE, show_header=False, show_lines=True)
table.add_column("Property", style="i")
table.add_column("Value")
if self.doc:
if include_full_doc:
table.add_row("Documentation", self.doc.full_doc)
else:
table.add_row("Description", self.doc.description)
# module_type_md = self.module.get_type_metadata()
if include_inputs:
inputs_table = create_table_from_field_schemas(
_add_required=True,
_add_default=True,
_show_header=True,
_constants=None,
**self.operation_details.inputs_schema,
)
table.add_row("Inputs", inputs_table)
if include_outputs:
outputs_table = create_table_from_field_schemas(
_add_required=False,
_add_default=False,
_show_header=True,
_constants=None,
**self.operation_details.outputs_schema,
)
table.add_row("Outputs", outputs_table)
if include_module_details:
table.add_row("Module type", self.module_type)
module_config = self.module.config.json(option=orjson.OPT_INDENT_2)
conf = Syntax(
module_config,
"json",
background_color="default",
)
table.add_row("Module config", conf)
module_type_md = KiaraModuleTypeInfo.create_from_type_class(
self.module_details.get_class() # type: ignore
)
desc = module_type_md.documentation.description
module_md = module_type_md.create_renderable(
include_doc=False, include_src=False, include_config_schema=False
)
m_md = Group(desc, module_md)
table.add_row("Module metadata", m_md)
if include_src:
table.add_row("Source code", module_type_md.process_src)
return table
prepare_job_config(self, kiara, inputs)
¶Source code in kiara/models/module/operation.py
def prepare_job_config(
self, kiara: "Kiara", inputs: Mapping[str, Any]
) -> JobConfig:
augmented_inputs = (
self.operation_details.get_operation_schema().augment_module_inputs(
inputs=inputs
)
)
module_inputs = self.operation_details.create_module_inputs(
inputs=augmented_inputs
)
job_config = kiara.job_registry.prepare_job_config(
manifest=self, inputs=module_inputs
)
return job_config
process_job_outputs(self, outputs)
¶Source code in kiara/models/module/operation.py
def process_job_outputs(self, outputs: ValueMap) -> ValueMap:
op_outputs = self.operation_details.create_operation_outputs(outputs=outputs)
value_set = ValueMapReadOnly(value_items=op_outputs, values_schema=self.outputs_schema) # type: ignore
return value_set
run(self, kiara, inputs)
¶Source code in kiara/models/module/operation.py
def run(self, kiara: "Kiara", inputs: Any) -> ValueMap:
logger.debug("run.operation", operation_id=self.operation_id)
job_config = self.prepare_job_config(kiara=kiara, inputs=inputs)
job_id = kiara.job_registry.execute_job(job_config=job_config)
outputs: ValueMap = kiara.job_registry.retrieve_result(job_id=job_id)
result = self.process_job_outputs(outputs=outputs)
return result
OperationConfig (KiaraModel)
pydantic-model
¶Source code in kiara/models/module/operation.py
class OperationConfig(KiaraModel):
doc: DocumentationMetadataModel = Field(
description="Documentation for this operation."
)
@validator("doc", pre=True)
def validate_doc(cls, value):
return DocumentationMetadataModel.create(value)
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return OPERATION_CONFIG_CATEOGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
@abc.abstractmethod
def retrieve_module_type(self, kiara: "Kiara") -> str:
pass
@abc.abstractmethod
def retrieve_module_config(self, kiara: "Kiara") -> Mapping[str, Any]:
pass
doc: DocumentationMetadataModel
pydantic-field
required
¶Documentation for this operation.
retrieve_module_config(self, kiara)
¶Source code in kiara/models/module/operation.py
@abc.abstractmethod
def retrieve_module_config(self, kiara: "Kiara") -> Mapping[str, Any]:
pass
retrieve_module_type(self, kiara)
¶Source code in kiara/models/module/operation.py
@abc.abstractmethod
def retrieve_module_type(self, kiara: "Kiara") -> str:
pass
validate_doc(value)
classmethod
¶Source code in kiara/models/module/operation.py
@validator("doc", pre=True)
def validate_doc(cls, value):
return DocumentationMetadataModel.create(value)
OperationDetails (KiaraModel)
pydantic-model
¶Source code in kiara/models/module/operation.py
class OperationDetails(KiaraModel):
@classmethod
def create_operation_details(cls, **details: Any):
if PYDANTIC_USE_CONSTRUCT:
result = cls.construct(**details)
else:
result = cls(**details)
return result
operation_id: str = Field(description="The id of the operation.")
is_internal_operation: bool = Field(
description="Whether this operation is mainly used kiara-internally. Helps to hide it in UIs (operation lists etc.).",
default=False,
)
def _retrieve_id(self) -> str:
return self.operation_id
def _retrieve_category_id(self) -> str:
return OPERATION_DETAILS_CATEOGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
@property
def inputs_schema(self) -> Mapping[str, ValueSchema]:
"""The input schema for this module."""
return self.get_operation_schema().inputs_schema
@property
def outputs_schema(self) -> Mapping[str, ValueSchema]:
"""The input schema for this module."""
return self.get_operation_schema().outputs_schema
def get_operation_schema(self) -> OperationSchema:
raise NotImplementedError()
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
raise NotImplementedError()
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
raise NotImplementedError()
inputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
property
readonly
¶The input schema for this module.
is_internal_operation: bool
pydantic-field
¶Whether this operation is mainly used kiara-internally. Helps to hide it in UIs (operation lists etc.).
operation_id: str
pydantic-field
required
¶The id of the operation.
outputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
property
readonly
¶The input schema for this module.
create_module_inputs(self, inputs)
¶Source code in kiara/models/module/operation.py
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
raise NotImplementedError()
create_operation_details(**details)
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def create_operation_details(cls, **details: Any):
if PYDANTIC_USE_CONSTRUCT:
result = cls.construct(**details)
else:
result = cls(**details)
return result
create_operation_outputs(self, outputs)
¶Source code in kiara/models/module/operation.py
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
raise NotImplementedError()
get_operation_schema(self)
¶Source code in kiara/models/module/operation.py
def get_operation_schema(self) -> OperationSchema:
raise NotImplementedError()
OperationGroupInfo (InfoModelGroup)
pydantic-model
¶Source code in kiara/models/module/operation.py
class OperationGroupInfo(InfoModelGroup):
@classmethod
def base_info_class(cls) -> Type[KiaraInfoModel]:
return OperationInfo
@classmethod
def create_from_operations(
cls, kiara: "Kiara", group_alias: Optional[str] = None, **items: Operation
) -> "OperationGroupInfo":
type_infos = {
k: OperationInfo.create_from_operation(kiara=kiara, operation=v)
for k, v in items.items()
}
data_types_info = cls.construct(group_alias=group_alias, type_infos=type_infos)
return data_types_info
type_name: Literal["operation_type"] = "operation_type"
type_infos: Mapping[str, OperationInfo] = Field(
description="The operation info instances for each type."
)
def create_renderable(self, **config: Any) -> RenderableType:
by_type = config.get("by_type", False)
if by_type:
return self._create_renderable_by_type(**config)
else:
return self._create_renderable_list(**config)
def _create_renderable_list(self, **config):
include_internal_operations = config.get("include_internal_operations", True)
full_doc = config.get("full_doc", False)
filter = config.get("filter", [])
table = Table(box=box.SIMPLE, show_header=True)
table.add_column("Id", no_wrap=True, style="i")
table.add_column("Type(s)", style="green")
table.add_column("Description")
for op_id, op_info in self.type_infos.items():
if (
not include_internal_operations
and op_info.operation.operation_details.is_internal_operation
):
continue
types = op_info.operation_types
try:
types.remove("custom_module")
except KeyError:
pass
desc_str = op_info.documentation.description
if full_doc:
desc = Markdown(op_info.documentation.full_doc)
else:
desc = Markdown(op_info.documentation.description)
if filter:
match = True
for f in filter:
if (
f.lower() not in op_id.lower()
and f.lower() not in desc_str.lower()
):
match = False
break
if match:
table.add_row(op_id, ", ".join(types), desc)
else:
table.add_row(op_id, ", ".join(types), desc)
return table
def _create_renderable_by_type(self, **config):
include_internal_operations = config.get("include_internal_operations", True)
full_doc = config.get("full_doc", False)
filter = config.get("filter", [])
by_type = {}
for op_id, op in self.type_infos.items():
if filter:
match = True
for f in filter:
if (
f.lower() not in op_id.lower()
and f.lower() not in op.documentation.description.lower()
):
match = False
break
if not match:
continue
for op_type in op.operation_types:
by_type.setdefault(op_type, {})[op_id] = op
table = Table(box=box.SIMPLE, show_header=True)
table.add_column("Type", no_wrap=True, style="b green")
table.add_column("Id", no_wrap=True, style="i")
if full_doc:
table.add_column("Documentation", no_wrap=False, style="i")
else:
table.add_column("Description", no_wrap=False, style="i")
for operation_name in sorted(by_type.keys()):
if operation_name == "custom_module":
continue
first_line_value = True
op_infos = by_type[operation_name]
for op_id in sorted(op_infos.keys()):
op_info: OperationInfo = op_infos[op_id]
if (
not include_internal_operations
and op_info.operation.operation_details.is_internal_operation
):
continue
if full_doc:
desc = Markdown(op_info.documentation.full_doc)
else:
desc = Markdown(op_info.documentation.description)
row = []
if first_line_value:
row.append(operation_name)
else:
row.append("")
row.append(op_id)
row.append(desc)
table.add_row(*row)
first_line_value = False
return table
type_infos: Mapping[str, kiara.models.module.operation.OperationInfo]
pydantic-field
required
¶The operation info instances for each type.
type_name: Literal['operation_type']
pydantic-field
¶base_info_class()
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def base_info_class(cls) -> Type[KiaraInfoModel]:
return OperationInfo
create_from_operations(kiara, group_alias=None, **items)
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def create_from_operations(
cls, kiara: "Kiara", group_alias: Optional[str] = None, **items: Operation
) -> "OperationGroupInfo":
type_infos = {
k: OperationInfo.create_from_operation(kiara=kiara, operation=v)
for k, v in items.items()
}
data_types_info = cls.construct(group_alias=group_alias, type_infos=type_infos)
return data_types_info
create_renderable(self, **config)
¶Source code in kiara/models/module/operation.py
def create_renderable(self, **config: Any) -> RenderableType:
by_type = config.get("by_type", False)
if by_type:
return self._create_renderable_by_type(**config)
else:
return self._create_renderable_list(**config)
OperationInfo (KiaraInfoModel)
pydantic-model
¶Source code in kiara/models/module/operation.py
class OperationInfo(KiaraInfoModel):
@classmethod
def create_from_operation(cls, kiara: "Kiara", operation: Operation):
module = operation.module
module_cls = module.__class__
authors_md = AuthorsMetadataModel.from_class(module_cls)
properties_md = ContextMetadataModel.from_class(module_cls)
op_types = kiara.operation_registry.find_all_operation_types(
operation_id=operation.operation_id
)
op_info = OperationInfo.construct(
type_name=operation.operation_id,
operation_types=op_types,
operation=operation,
documentation=operation.doc,
authors=authors_md,
context=properties_md,
)
return op_info
@classmethod
def category_name(cls) -> str:
return "operation"
operation: Operation = Field(description="The operation instance.")
operation_types: Set[str] = Field(
description="The operation types this operation belongs to."
)
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable(**config))
table.add_row("Context", self.context.create_renderable(**config))
table.add_row("Operation details", self.operation.create_renderable(**config))
return table
operation: Operation
pydantic-field
required
¶The operation instance.
operation_types: Set[str]
pydantic-field
required
¶The operation types this operation belongs to.
category_name()
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def category_name(cls) -> str:
return "operation"
create_from_operation(kiara, operation)
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def create_from_operation(cls, kiara: "Kiara", operation: Operation):
module = operation.module
module_cls = module.__class__
authors_md = AuthorsMetadataModel.from_class(module_cls)
properties_md = ContextMetadataModel.from_class(module_cls)
op_types = kiara.operation_registry.find_all_operation_types(
operation_id=operation.operation_id
)
op_info = OperationInfo.construct(
type_name=operation.operation_id,
operation_types=op_types,
operation=operation,
documentation=operation.doc,
authors=authors_md,
context=properties_md,
)
return op_info
create_renderable(self, **config)
¶Source code in kiara/models/module/operation.py
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable(**config))
table.add_row("Context", self.context.create_renderable(**config))
table.add_row("Operation details", self.operation.create_renderable(**config))
return table
OperationSchema (InputOutputObject)
¶Source code in kiara/models/module/operation.py
class OperationSchema(InputOutputObject):
def __init__(
self, alias: str, inputs_schema: ValueSetSchema, outputs_schema: ValueSetSchema
):
allow_empty_inputs = True
allow_empty_outputs = True
self._inputs_schema_static: ValueSetSchema = inputs_schema
self._outputs_schema_static: ValueSetSchema = outputs_schema
super().__init__(
alias=alias,
allow_empty_inputs_schema=allow_empty_inputs,
allow_empty_outputs_schema=allow_empty_outputs,
)
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return self._inputs_schema_static
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return self._outputs_schema_static
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/models/module/operation.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return self._inputs_schema_static
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/models/module/operation.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return self._outputs_schema_static
OperationTypeClassesInfo (TypeInfoModelGroup)
pydantic-model
¶Source code in kiara/models/module/operation.py
class OperationTypeClassesInfo(TypeInfoModelGroup):
@classmethod
def base_info_class(cls) -> Type[KiaraTypeInfoModel]:
return OperationTypeInfo
type_name: Literal["operation_type"] = "operation_type"
type_infos: Mapping[str, OperationTypeInfo] = Field(
description="The operation info instances for each type."
)
type_infos: Mapping[str, kiara.models.module.operation.OperationTypeInfo]
pydantic-field
required
¶The operation info instances for each type.
type_name: Literal['operation_type']
pydantic-field
¶base_info_class()
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def base_info_class(cls) -> Type[KiaraTypeInfoModel]:
return OperationTypeInfo
OperationTypeInfo (KiaraTypeInfoModel)
pydantic-model
¶Source code in kiara/models/module/operation.py
class OperationTypeInfo(KiaraTypeInfoModel):
@classmethod
def create_from_type_class(
cls, type_cls: Type["OperationType"]
) -> "OperationTypeInfo":
authors_md = AuthorsMetadataModel.from_class(type_cls)
doc = DocumentationMetadataModel.from_class_doc(type_cls)
python_class = PythonClass.from_class(type_cls)
properties_md = ContextMetadataModel.from_class(type_cls)
return OperationTypeInfo.construct(
**{
"type_name": type_cls._operation_type_name, # type: ignore
"documentation": doc,
"authors": authors_md,
"context": properties_md,
"python_class": python_class,
}
)
@classmethod
def base_class(self) -> Type["OperationType"]:
from kiara.operations import OperationType
return OperationType
@classmethod
def category_name(cls) -> str:
return "operation_type"
def _retrieve_id(self) -> str:
return self.type_name
def _retrieve_category_id(self) -> str:
return OPERATION_TYPE_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.type_name
base_class()
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def base_class(self) -> Type["OperationType"]:
from kiara.operations import OperationType
return OperationType
category_name()
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def category_name(cls) -> str:
return "operation_type"
create_from_type_class(type_cls)
classmethod
¶Source code in kiara/models/module/operation.py
@classmethod
def create_from_type_class(
cls, type_cls: Type["OperationType"]
) -> "OperationTypeInfo":
authors_md = AuthorsMetadataModel.from_class(type_cls)
doc = DocumentationMetadataModel.from_class_doc(type_cls)
python_class = PythonClass.from_class(type_cls)
properties_md = ContextMetadataModel.from_class(type_cls)
return OperationTypeInfo.construct(
**{
"type_name": type_cls._operation_type_name, # type: ignore
"documentation": doc,
"authors": authors_md,
"context": properties_md,
"python_class": python_class,
}
)
PipelineOperationConfig (OperationConfig)
pydantic-model
¶Source code in kiara/models/module/operation.py
class PipelineOperationConfig(OperationConfig):
pipeline_id: str = Field(description="The pipeline id.")
pipeline_config: Mapping[str, Any] = Field(description="The pipeline config data.")
module_map: Dict[str, Any] = Field(
description="A lookup map to resolves module names to operations.",
default_factory=dict,
)
@validator("pipeline_config")
def validate_pipeline_config(cls, value):
# TODO
assert isinstance(value, Mapping)
assert "steps" in value.keys()
return value
def retrieve_module_type(self, kiara: "Kiara") -> str:
return "pipeline"
def retrieve_module_config(self, kiara: "Kiara") -> Mapping[str, Any]:
# using _from_config here because otherwise we'd enter an infinite loop
pipeline_config = PipelineConfig._from_config(
pipeline_id=self.pipeline_id,
data=self.pipeline_config,
kiara=kiara,
module_map=self.module_map,
)
return pipeline_config.dict()
@property
def required_module_types(self) -> Iterable[str]:
return [step["module_type"] for step in self.pipeline_config["steps"]]
def __repr__(self):
return f"{self.__class__.__name__}(pipeline_id={self.pipeline_id} required_modules={list(self.required_module_types)} model_id={self.model_id}, category={self.category_id}, fields=[{', '.join(self.__fields__.keys())}])"
module_map: Dict[str, Any]
pydantic-field
¶A lookup map to resolves module names to operations.
pipeline_config: Mapping[str, Any]
pydantic-field
required
¶The pipeline config data.
pipeline_id: str
pydantic-field
required
¶The pipeline id.
required_module_types: Iterable[str]
property
readonly
¶retrieve_module_config(self, kiara)
¶Source code in kiara/models/module/operation.py
def retrieve_module_config(self, kiara: "Kiara") -> Mapping[str, Any]:
# using _from_config here because otherwise we'd enter an infinite loop
pipeline_config = PipelineConfig._from_config(
pipeline_id=self.pipeline_id,
data=self.pipeline_config,
kiara=kiara,
module_map=self.module_map,
)
return pipeline_config.dict()
retrieve_module_type(self, kiara)
¶Source code in kiara/models/module/operation.py
def retrieve_module_type(self, kiara: "Kiara") -> str:
return "pipeline"
validate_pipeline_config(value)
classmethod
¶Source code in kiara/models/module/operation.py
@validator("pipeline_config")
def validate_pipeline_config(cls, value):
# TODO
assert isinstance(value, Mapping)
assert "steps" in value.keys()
return value
persistence
¶
ByteProvisioningStrategy (Enum)
¶
BytesAliasStructure (BaseModel)
pydantic-model
¶Source code in kiara/models/module/persistence.py
class BytesAliasStructure(BaseModel):
data_type: str = Field(description="The data type.")
data_type_config: Mapping[str, Any] = Field(description="The data type config.")
chunk_id_map: Mapping[str, List[str]] = Field(
description="References to byte arrays, Keys are field names, values are a list of hash-ids that the data is composed of.",
default_factory=dict,
)
chunk_id_map: Mapping[str, List[str]]
pydantic-field
¶References to byte arrays, Keys are field names, values are a list of hash-ids that the data is composed of.
data_type: str
pydantic-field
required
¶The data type.
data_type_config: Mapping[str, Any]
pydantic-field
required
¶The data type config.
BytesStructure (BaseModel)
pydantic-model
¶A data structure that
Source code in kiara/models/module/persistence.py
class BytesStructure(BaseModel):
"""A data structure that"""
data_type: str = Field(description="The data type.")
data_type_config: Mapping[str, Any] = Field(description="The data type config.")
chunk_map: Mapping[str, List[Union[str, bytes]]] = Field(
description="References to byte arrays, Keys are field names, values are a list of hash-ids that the data is composed of.",
default_factory=dict,
)
def provision_as_folder(self, copy_files: bool = False) -> Path:
pass
chunk_map: Mapping[str, List[Union[str, bytes]]]
pydantic-field
¶References to byte arrays, Keys are field names, values are a list of hash-ids that the data is composed of.
data_type: str
pydantic-field
required
¶The data type.
data_type_config: Mapping[str, Any]
pydantic-field
required
¶The data type config.
provision_as_folder(self, copy_files=False)
¶Source code in kiara/models/module/persistence.py
def provision_as_folder(self, copy_files: bool = False) -> Path:
pass
LoadConfig (Manifest)
pydantic-model
¶Source code in kiara/models/module/persistence.py
class LoadConfig(Manifest):
provisioning_strategy: ByteProvisioningStrategy = Field(
description="In what form the serialized bytes are returned.",
default=ByteProvisioningStrategy.INLINE,
)
inputs: Mapping[str, str] = Field(
description="A map translating from input field name to alias (key) in bytes_structure."
)
output_name: str = Field(
description="The name of the field that contains the persisted value details."
)
bytes_map: Optional[BytesAliasStructure] = Field(
description="References to the byte chunks for the inputs.", default=None
)
inline_data: Optional[Any] = Field(description="Inline values.", default=None)
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
def __repr__(self):
return f"{self.__class__.__name__}(module_type={self.module_type}, output_name={self.output_name})"
def __str__(self):
return self.__repr__()
bytes_map: BytesAliasStructure
pydantic-field
¶References to the byte chunks for the inputs.
inline_data: Any
pydantic-field
¶Inline values.
inputs: Mapping[str, str]
pydantic-field
required
¶A map translating from input field name to alias (key) in bytes_structure.
output_name: str
pydantic-field
required
¶The name of the field that contains the persisted value details.
provisioning_strategy: ByteProvisioningStrategy
pydantic-field
¶In what form the serialized bytes are returned.
pipeline
special
¶
PipelineConfig (KiaraModuleConfig)
pydantic-model
¶A class to hold the configuration for a [PipelineModule][kiara.pipeline.module.PipelineModule].
If you want to control the pipeline input and output names, you need to have to provide a map that uses the autogenerated field name ([step_id]__[alias] -- 2 underscores!!) as key, and the desired field name as value. The reason that schema for the autogenerated field names exist is that it's hard to ensure the uniqueness of each field; some steps can have the same input field names, but will need different input values. In some cases, some inputs of different steps need the same input. Those sorts of things. So, to make sure that we always use the right values, I chose to implement a conservative default approach, accepting that in some cases the user will be prompted for duplicate inputs for the same value.
To remedy that, the pipeline creator has the option to manually specify a mapping to rename some or all of the input/output fields.
Further, because in a lot of cases there won't be any overlapping fields, the creator can specify auto,
in which case Kiara will automatically create a mapping that tries to map autogenerated field names
to the shortest possible names for each case.
Examples:
Configuration for a pipeline module that functions as a nand logic gate (in Python):
and_step = PipelineStepConfig(module_type="and", step_id="and")
not_step = PipelineStepConfig(module_type="not", step_id="not", input_links={"a": ["and.y"]}
nand_p_conf = PipelineConfig(doc="Returns 'False' if both inputs are 'True'.",
steps=[and_step, not_step],
input_aliases={
"and__a": "a",
"and__b": "b"
},
output_aliases={
"not__y": "y"
}}
Or, the same thing in json:
{
"module_type_name": "nand",
"doc": "Returns 'False' if both inputs are 'True'.",
"steps": [
{
"module_type": "and",
"step_id": "and"
},
{
"module_type": "not",
"step_id": "not",
"input_links": {
"a": "and.y"
}
}
],
"input_aliases": {
"and__a": "a",
"and__b": "b"
},
"output_aliases": {
"not__y": "y"
}
}
Source code in kiara/models/module/pipeline/__init__.py
class PipelineConfig(KiaraModuleConfig):
"""A class to hold the configuration for a [PipelineModule][kiara.pipeline.module.PipelineModule].
If you want to control the pipeline input and output names, you need to have to provide a map that uses the
autogenerated field name ([step_id]__[alias] -- 2 underscores!!) as key, and the desired field name
as value. The reason that schema for the autogenerated field names exist is that it's hard to ensure
the uniqueness of each field; some steps can have the same input field names, but will need different input
values. In some cases, some inputs of different steps need the same input. Those sorts of things.
So, to make sure that we always use the right values, I chose to implement a conservative default approach,
accepting that in some cases the user will be prompted for duplicate inputs for the same value.
To remedy that, the pipeline creator has the option to manually specify a mapping to rename some or all of
the input/output fields.
Further, because in a lot of cases there won't be any overlapping fields, the creator can specify ``auto``,
in which case *Kiara* will automatically create a mapping that tries to map autogenerated field names
to the shortest possible names for each case.
Examples:
Configuration for a pipeline module that functions as a ``nand`` logic gate (in Python):
``` python
and_step = PipelineStepConfig(module_type="and", step_id="and")
not_step = PipelineStepConfig(module_type="not", step_id="not", input_links={"a": ["and.y"]}
nand_p_conf = PipelineConfig(doc="Returns 'False' if both inputs are 'True'.",
steps=[and_step, not_step],
input_aliases={
"and__a": "a",
"and__b": "b"
},
output_aliases={
"not__y": "y"
}}
```
Or, the same thing in json:
``` json
{
"module_type_name": "nand",
"doc": "Returns 'False' if both inputs are 'True'.",
"steps": [
{
"module_type": "and",
"step_id": "and"
},
{
"module_type": "not",
"step_id": "not",
"input_links": {
"a": "and.y"
}
}
],
"input_aliases": {
"and__a": "a",
"and__b": "b"
},
"output_aliases": {
"not__y": "y"
}
}
```
"""
@classmethod
def from_file(
cls,
path: str,
kiara: Optional["Kiara"] = None,
# module_map: Optional[Mapping[str, Any]] = None,
):
data = get_data_from_file(path)
pipeline_id = data.pop("pipeline_id", None)
if pipeline_id is None:
pipeline_id = os.path.basename(path)
return cls.from_config(pipeline_id=pipeline_id, data=data, kiara=kiara)
@classmethod
def from_config(
cls,
pipeline_id: str,
data: Mapping[str, Any],
kiara: Optional["Kiara"] = None,
# module_map: Optional[Mapping[str, Any]] = None,
):
if kiara is None:
from kiara.context import Kiara
kiara = Kiara.instance()
if not kiara.operation_registry.is_initialized:
kiara.operation_registry.operations # noqa
return cls._from_config(pipeline_id=pipeline_id, data=data, kiara=kiara)
@classmethod
def _from_config(
cls,
pipeline_id: str,
data: Mapping[str, Any],
kiara: "Kiara",
module_map: Optional[Mapping[str, Any]] = None,
):
data = dict(data)
steps = data.pop("steps")
steps = PipelineStep.create_steps(*steps, kiara=kiara, module_map=module_map)
data["steps"] = steps
if not data.get("input_aliases"):
data["input_aliases"] = create_input_alias_map(steps)
if not data.get("output_aliases"):
data["output_aliases"] = create_output_alias_map(steps)
result = cls(pipeline_id=pipeline_id, **data)
return result
class Config:
extra = Extra.allow
validate_assignment = True
pipeline_id: str = Field(description="The name of this pipeline.")
steps: List[PipelineStep] = Field(
description="A list of steps/modules of this pipeline, and their connections.",
)
input_aliases: Dict[str, str] = Field(
description="A map of input aliases, with the calculated (<step_id>__<input_name> -- double underscore!) name as key, and a string (the resulting workflow input alias) as value. Check the documentation for the config class for which marker strings can be used to automatically create this map if possible.",
)
output_aliases: Dict[str, str] = Field(
description="A map of output aliases, with the calculated (<step_id>__<output_name> -- double underscore!) name as key, and a string (the resulting workflow output alias) as value. Check the documentation for the config class for which marker strings can be used to automatically create this map if possible.",
)
documentation: str = Field(
default="-- n/a --", description="Documentation about what the pipeline does."
)
context: Dict[str, Any] = Field(
default_factory=dict, description="Metadata for this workflow."
)
_structure: Optional["PipelineStructure"] = PrivateAttr(default=None)
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return PIPELINE_CONFIG_TYPE_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
@validator("steps", pre=True)
def _validate_steps(cls, v):
if not v:
raise ValueError(f"Invalid type for 'steps' value: {type(v)}")
steps = []
for step in v:
if not step:
raise ValueError("No step data provided.")
if isinstance(step, PipelineStep):
steps.append(step)
elif isinstance(step, Mapping):
steps.append(PipelineStep(**step))
else:
raise TypeError(step)
return steps
@property
def structure(self) -> "PipelineStructure":
if self._structure is not None:
return self._structure
from kiara.models.module.pipeline.structure import PipelineStructure
structure = PipelineStructure(pipeline_config=self)
return structure
def create_renderable(self, **config: Any) -> RenderableType:
return create_table_from_model_object(self, exclude_fields={"steps"})
# def create_input_alias_map(self) -> Dict[str, str]:
#
# aliases: Dict[str, List[str]] = {}
# for step in self.steps:
# field_names = step.module.input_names
# for field_name in field_names:
# aliases.setdefault(field_name, []).append(step.step_id)
#
# result: Dict[str, str] = {}
# for field_name, step_ids in aliases.items():
# for step_id in step_ids:
# generated = generate_pipeline_endpoint_name(step_id, field_name)
# result[generated] = generated
#
# return result
#
# def create_output_alias_map(self) -> Dict[str, str]:
#
# aliases: Dict[str, List[str]] = {}
# for step in self.steps:
# field_names = step.module.input_names
# for field_name in field_names:
# aliases.setdefault(field_name, []).append(step.step_id)
#
# result: Dict[str, str] = {}
# for field_name, step_ids in aliases.items():
# for step_id in step_ids:
# generated = generate_pipeline_endpoint_name(step_id, field_name)
# result[generated] = generated
#
# return result
context: Dict[str, Any]
pydantic-field
¶Metadata for this workflow.
documentation: str
pydantic-field
¶Documentation about what the pipeline does.
input_aliases: Dict[str, str]
pydantic-field
required
¶A map of input aliases, with the calculated (
output_aliases: Dict[str, str]
pydantic-field
required
¶A map of output aliases, with the calculated (
pipeline_id: str
pydantic-field
required
¶The name of this pipeline.
steps: List[kiara.models.module.pipeline.PipelineStep]
pydantic-field
required
¶A list of steps/modules of this pipeline, and their connections.
structure: PipelineStructure
property
readonly
¶
Config
¶create_renderable(self, **config)
¶Source code in kiara/models/module/pipeline/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
return create_table_from_model_object(self, exclude_fields={"steps"})
from_config(pipeline_id, data, kiara=None)
classmethod
¶Source code in kiara/models/module/pipeline/__init__.py
@classmethod
def from_config(
cls,
pipeline_id: str,
data: Mapping[str, Any],
kiara: Optional["Kiara"] = None,
# module_map: Optional[Mapping[str, Any]] = None,
):
if kiara is None:
from kiara.context import Kiara
kiara = Kiara.instance()
if not kiara.operation_registry.is_initialized:
kiara.operation_registry.operations # noqa
return cls._from_config(pipeline_id=pipeline_id, data=data, kiara=kiara)
from_file(path, kiara=None)
classmethod
¶Source code in kiara/models/module/pipeline/__init__.py
@classmethod
def from_file(
cls,
path: str,
kiara: Optional["Kiara"] = None,
# module_map: Optional[Mapping[str, Any]] = None,
):
data = get_data_from_file(path)
pipeline_id = data.pop("pipeline_id", None)
if pipeline_id is None:
pipeline_id = os.path.basename(path)
return cls.from_config(pipeline_id=pipeline_id, data=data, kiara=kiara)
PipelineStep (Manifest)
pydantic-model
¶A step within a pipeline-structure, includes information about it's connection(s) and other metadata.
Source code in kiara/models/module/pipeline/__init__.py
class PipelineStep(Manifest):
"""A step within a pipeline-structure, includes information about it's connection(s) and other metadata."""
class Config:
validate_assignment = True
extra = Extra.forbid
@classmethod
def create_steps(
cls,
*steps: Mapping[str, Any],
kiara: "Kiara",
module_map: Optional[Mapping[str, Any]] = None,
) -> List["PipelineStep"]:
if module_map is None:
module_map = {}
else:
module_map = dict(module_map)
if kiara.operation_registry.is_initialized:
for op_id, op in kiara.operation_registry.operations.items():
module_map[op_id] = {
"module_type": op.module_type,
"module_config": op.module_config,
}
result: List[PipelineStep] = []
for step in steps:
module_type = step.get("module_type", None)
if not module_type:
raise ValueError("Can't create step, no 'module_type' specified.")
module_config = step.get("module_config", {})
if module_type not in kiara.module_type_names:
if module_type in module_map.keys():
resolved_module_type = module_map[module_type]["module_type"]
resolved_module_config = module_map[module_type]["module_config"]
manifest = kiara.create_manifest(
module_or_operation=resolved_module_type,
config=resolved_module_config,
)
else:
raise Exception(f"Can't resolve module type: {module_type}")
else:
manifest = kiara.create_manifest(
module_or_operation=module_type, config=module_config
)
resolved_module_type = module_type
resolved_module_config = module_config
module = kiara.create_module(manifest=manifest)
step_id = step.get("step_id", None)
if not step_id:
raise ValueError("Can't create step, no 'step_id' specified.")
input_links = {}
for input_field, sources in step.get("input_links", {}).items():
if isinstance(sources, str):
sources = [sources]
input_links[input_field] = sources
# TODO: do we really need the deepcopy here?
_s = PipelineStep(
step_id=step_id,
module_type=resolved_module_type,
module_config=dict(resolved_module_config),
input_links=input_links, # type: ignore
module_details=KiaraModuleClass.from_module(module=module),
)
_s._module = module
result.append(_s)
return result
@validator("step_id")
def _validate_step_id(cls, v):
assert isinstance(v, str)
if "." in v:
raise ValueError("Step ids can't contain '.' characters.")
return v
step_id: str = Field(
description="Locally unique id (within a pipeline) of this step."
)
module_type: str = Field(description="The module type.")
module_config: Dict[str, Any] = Field(
description="The module config.", default_factory=dict
)
# required: bool = Field(
# description="Whether this step is required within the workflow.\n\nIn some cases, when none of the pipeline outputs have a required input that connects to a step, then it is not necessary for this step to have been executed, even if it is placed before a step in the execution hierarchy. This also means that the pipeline inputs that are connected to this step might not be required.",
# default=True,
# )
# processing_stage: Optional[int] = Field(
# default=None,
# description="The stage number this step is executed within the pipeline.",
# )
input_links: Mapping[str, List[StepValueAddress]] = Field(
description="The links that connect to inputs of the module.",
default_factory=list,
)
module_details: KiaraModuleClass = Field(
description="The class of the underlying module."
)
_module: Optional["KiaraModule"] = PrivateAttr(default=None)
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return PIPELINE_STEP_TYPE_CATEGORY_ID
@root_validator(pre=True)
def create_step_id(cls, values):
if "module_type" not in values:
raise ValueError("No 'module_type' specified.")
if "step_id" not in values or not values["step_id"]:
values["step_id"] = slugify(values["module_type"])
return values
@validator("step_id")
def ensure_valid_id(cls, v):
# TODO: check with regex
if "." in v or " " in v:
raise ValueError(
f"Step id can't contain special characters or whitespaces: {v}"
)
return v
@validator("module_config", pre=True)
def ensure_dict(cls, v):
if v is None:
v = {}
return v
@validator("input_links", pre=True)
def ensure_input_links_valid(cls, v):
if v is None:
v = {}
result = {}
for input_name, output in v.items():
input_links = ensure_step_value_addresses(
default_field_name=input_name, link=output
)
result[input_name] = input_links
return result
@property
def module(self) -> "KiaraModule":
if self._module is None:
m_cls = self.module_details.get_class()
self._module = m_cls(module_config=self.module_config)
return self._module
def __repr__(self):
return f"{self.__class__.__name__}(step_id={self.step_id} module_type={self.module_type})"
def __str__(self):
return f"step: {self.step_id} (module: {self.module_type})"
input_links: Mapping[str, List[kiara.models.module.pipeline.value_refs.StepValueAddress]]
pydantic-field
¶The links that connect to inputs of the module.
module: KiaraModule
property
readonly
¶module_details: KiaraModuleClass
pydantic-field
required
¶The class of the underlying module.
step_id: str
pydantic-field
required
¶Locally unique id (within a pipeline) of this step.
Config
¶create_step_id(values)
classmethod
¶Source code in kiara/models/module/pipeline/__init__.py
@root_validator(pre=True)
def create_step_id(cls, values):
if "module_type" not in values:
raise ValueError("No 'module_type' specified.")
if "step_id" not in values or not values["step_id"]:
values["step_id"] = slugify(values["module_type"])
return values
create_steps(*steps, *, kiara, module_map=None)
classmethod
¶Source code in kiara/models/module/pipeline/__init__.py
@classmethod
def create_steps(
cls,
*steps: Mapping[str, Any],
kiara: "Kiara",
module_map: Optional[Mapping[str, Any]] = None,
) -> List["PipelineStep"]:
if module_map is None:
module_map = {}
else:
module_map = dict(module_map)
if kiara.operation_registry.is_initialized:
for op_id, op in kiara.operation_registry.operations.items():
module_map[op_id] = {
"module_type": op.module_type,
"module_config": op.module_config,
}
result: List[PipelineStep] = []
for step in steps:
module_type = step.get("module_type", None)
if not module_type:
raise ValueError("Can't create step, no 'module_type' specified.")
module_config = step.get("module_config", {})
if module_type not in kiara.module_type_names:
if module_type in module_map.keys():
resolved_module_type = module_map[module_type]["module_type"]
resolved_module_config = module_map[module_type]["module_config"]
manifest = kiara.create_manifest(
module_or_operation=resolved_module_type,
config=resolved_module_config,
)
else:
raise Exception(f"Can't resolve module type: {module_type}")
else:
manifest = kiara.create_manifest(
module_or_operation=module_type, config=module_config
)
resolved_module_type = module_type
resolved_module_config = module_config
module = kiara.create_module(manifest=manifest)
step_id = step.get("step_id", None)
if not step_id:
raise ValueError("Can't create step, no 'step_id' specified.")
input_links = {}
for input_field, sources in step.get("input_links", {}).items():
if isinstance(sources, str):
sources = [sources]
input_links[input_field] = sources
# TODO: do we really need the deepcopy here?
_s = PipelineStep(
step_id=step_id,
module_type=resolved_module_type,
module_config=dict(resolved_module_config),
input_links=input_links, # type: ignore
module_details=KiaraModuleClass.from_module(module=module),
)
_s._module = module
result.append(_s)
return result
ensure_dict(v)
classmethod
¶Source code in kiara/models/module/pipeline/__init__.py
@validator("module_config", pre=True)
def ensure_dict(cls, v):
if v is None:
v = {}
return v
ensure_input_links_valid(v)
classmethod
¶Source code in kiara/models/module/pipeline/__init__.py
@validator("input_links", pre=True)
def ensure_input_links_valid(cls, v):
if v is None:
v = {}
result = {}
for input_name, output in v.items():
input_links = ensure_step_value_addresses(
default_field_name=input_name, link=output
)
result[input_name] = input_links
return result
ensure_valid_id(v)
classmethod
¶Source code in kiara/models/module/pipeline/__init__.py
@validator("step_id")
def ensure_valid_id(cls, v):
# TODO: check with regex
if "." in v or " " in v:
raise ValueError(
f"Step id can't contain special characters or whitespaces: {v}"
)
return v
StepStatus (Enum)
¶Enum to describe the state of a workflow.
Source code in kiara/models/module/pipeline/__init__.py
class StepStatus(Enum):
"""Enum to describe the state of a workflow."""
INPUTS_INVALID = "inputs_invalid"
INPUTS_READY = "inputs_ready"
RESULTS_READY = "results_ready"
create_input_alias_map(steps)
¶Source code in kiara/models/module/pipeline/__init__.py
def create_input_alias_map(steps: Iterable[PipelineStep]) -> Dict[str, str]:
aliases: Dict[str, str] = {}
for step in steps:
field_names = step.module.input_names
for field_name in field_names:
alias = generate_pipeline_endpoint_name(
step_id=step.step_id, value_name=field_name
)
assert alias not in aliases.keys()
aliases[f"{step.step_id}.{field_name}"] = alias
return aliases
create_output_alias_map(steps)
¶Source code in kiara/models/module/pipeline/__init__.py
def create_output_alias_map(steps: Iterable[PipelineStep]) -> Dict[str, str]:
aliases: Dict[str, str] = {}
for step in steps:
field_names = step.module.output_names
for field_name in field_names:
alias = generate_pipeline_endpoint_name(
step_id=step.step_id, value_name=field_name
)
assert alias not in aliases.keys()
aliases[f"{step.step_id}.{field_name}"] = alias
return aliases
generate_pipeline_endpoint_name(step_id, value_name)
¶Source code in kiara/models/module/pipeline/__init__.py
def generate_pipeline_endpoint_name(step_id: str, value_name: str):
return f"{step_id}__{value_name}"
controller
¶logger
¶
PipelineController (PipelineListener)
¶Source code in kiara/models/module/pipeline/controller.py
class PipelineController(PipelineListener):
pass
SinglePipelineBatchController (SinglePipelineController)
¶A [PipelineController][kiara.models.modules.pipeline.controller.PipelineController] that executes all pipeline steps non-interactively.
This is the default implementation of a PipelineController, and probably the most simple implementation of one.
It waits until all inputs are set, after which it executes all pipeline steps in the required order.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
pipeline |
Pipeline |
the pipeline to control |
required |
auto_process |
bool |
whether to automatically start processing the pipeline as soon as the input set is valid |
True |
Source code in kiara/models/module/pipeline/controller.py
class SinglePipelineBatchController(SinglePipelineController):
"""A [PipelineController][kiara.models.modules.pipeline.controller.PipelineController] that executes all pipeline steps non-interactively.
This is the default implementation of a ``PipelineController``, and probably the most simple implementation of one.
It waits until all inputs are set, after which it executes all pipeline steps in the required order.
Arguments:
pipeline: the pipeline to control
auto_process: whether to automatically start processing the pipeline as soon as the input set is valid
"""
def __init__(
self,
pipeline: Pipeline,
job_registry: JobRegistry,
auto_process: bool = True,
):
self._auto_process: bool = auto_process
self._is_running: bool = False
super().__init__(pipeline=pipeline, job_registry=job_registry)
@property
def auto_process(self) -> bool:
return self._auto_process
@auto_process.setter
def auto_process(self, auto_process: bool):
self._auto_process = auto_process
def process_pipeline(self):
if self._is_running:
logger.debug(
"ignore.pipeline_process",
reason="Pipeline already running.",
pipeline_id=self.pipeline.pipeline_id,
)
raise Exception("Pipeline already running.")
logger.debug("execute.pipeline", pipeline_id=self.pipeline.pipeline_id)
self._is_running = True
try:
for idx, stage in enumerate(
self.pipeline.structure.processing_stages, start=1
):
logger.debug(
"execute.pipeline.stage",
pipeline_id=self.pipeline.pipeline_id,
stage=idx,
)
job_ids = {}
for step_id in stage:
logger.debug(
"execute.pipeline.step",
pipeline_id=self.pipeline.pipeline_id,
step_id=step_id,
)
try:
job_id = self.process_step(step_id)
job_ids[step_id] = job_id
except Exception as e:
# TODO: cancel running jobs?
if is_debug():
import traceback
traceback.print_exc()
logger.error(
"error.processing.pipeline",
step_id=step_id,
pipeline_id=self.pipeline.pipeline_id,
error=e,
)
return False
self.set_processing_results(job_ids=job_ids)
logger.debug(
"execute_finished.pipeline.stage",
pipeline_id=self.pipeline.pipeline_id,
stage=idx,
)
finally:
self._is_running = False
logger.debug("execute_finished.pipeline", pipeline_id=self.pipeline.pipeline_id)
auto_process: bool
property
writable
¶process_pipeline(self)
¶Source code in kiara/models/module/pipeline/controller.py
def process_pipeline(self):
if self._is_running:
logger.debug(
"ignore.pipeline_process",
reason="Pipeline already running.",
pipeline_id=self.pipeline.pipeline_id,
)
raise Exception("Pipeline already running.")
logger.debug("execute.pipeline", pipeline_id=self.pipeline.pipeline_id)
self._is_running = True
try:
for idx, stage in enumerate(
self.pipeline.structure.processing_stages, start=1
):
logger.debug(
"execute.pipeline.stage",
pipeline_id=self.pipeline.pipeline_id,
stage=idx,
)
job_ids = {}
for step_id in stage:
logger.debug(
"execute.pipeline.step",
pipeline_id=self.pipeline.pipeline_id,
step_id=step_id,
)
try:
job_id = self.process_step(step_id)
job_ids[step_id] = job_id
except Exception as e:
# TODO: cancel running jobs?
if is_debug():
import traceback
traceback.print_exc()
logger.error(
"error.processing.pipeline",
step_id=step_id,
pipeline_id=self.pipeline.pipeline_id,
error=e,
)
return False
self.set_processing_results(job_ids=job_ids)
logger.debug(
"execute_finished.pipeline.stage",
pipeline_id=self.pipeline.pipeline_id,
stage=idx,
)
finally:
self._is_running = False
logger.debug("execute_finished.pipeline", pipeline_id=self.pipeline.pipeline_id)
SinglePipelineController (PipelineController)
¶Source code in kiara/models/module/pipeline/controller.py
class SinglePipelineController(PipelineController):
def __init__(self, pipeline: Pipeline, job_registry: JobRegistry):
self._pipeline: Pipeline = pipeline
self._job_registry: JobRegistry = job_registry
self._pipeline.add_listener(self)
self._pipeline_details: Optional[PipelineDetails] = None
@property
def pipeline(self) -> Pipeline:
return self._pipeline
def current_pipeline_state(self) -> PipelineDetails:
if self._pipeline_details is None:
self._pipeline_details = self.pipeline.get_pipeline_details()
return self._pipeline_details
def can_be_processed(self, step_id: str) -> bool:
"""Check whether the step with the provided id is ready to be processed."""
pipeline_state = self.current_pipeline_state()
step_state = pipeline_state.step_states[step_id]
return not step_state.invalid_details
def can_be_skipped(self, step_id: str) -> bool:
"""Check whether the processing of a step can be skipped."""
required = self.pipeline.structure.step_is_required(step_id=step_id)
if required:
required = self.can_be_processed(step_id)
return required
def _pipeline_event_occurred(self, event: PipelineEvent):
self._pipeline_details = None
def set_processing_results(self, job_ids: Mapping[str, uuid.UUID]):
self._job_registry.wait_for(*job_ids.values())
combined_outputs = {}
for step_id, job_id in job_ids.items():
record = self._job_registry.get_job_record_in_session(job_id=job_id)
combined_outputs[step_id] = record.outputs
self.pipeline.set_multiple_step_outputs(
changed_outputs=combined_outputs, notify_listeners=True
)
def pipeline_is_ready(self) -> bool:
"""Return whether the pipeline is ready to be processed.
A ``True`` result means that all pipeline inputs are set with valid values, and therefore every step within the
pipeline can be processed.
Returns:
whether the pipeline can be processed as a whole (``True``) or not (``False``)
"""
pipeline_inputs = self.pipeline._all_values.get_alias("pipeline.inputs")
assert pipeline_inputs is not None
return pipeline_inputs.all_items_valid
def process_step(self, step_id: str, wait: bool = False) -> uuid.UUID:
"""Kick off processing for the step with the provided id.
Arguments:
step_id: the id of the step that should be started
"""
job_config = self.pipeline.create_job_config_for_step(step_id)
job_id = self._job_registry.execute_job(job_config=job_config)
# job_id = self._processor.create_job(job_config=job_config)
# self._processor.queue_job(job_id=job_id)
if wait:
self._job_registry.wait_for(job_id)
return job_id
pipeline: Pipeline
property
readonly
¶can_be_processed(self, step_id)
¶Check whether the step with the provided id is ready to be processed.
Source code in kiara/models/module/pipeline/controller.py
def can_be_processed(self, step_id: str) -> bool:
"""Check whether the step with the provided id is ready to be processed."""
pipeline_state = self.current_pipeline_state()
step_state = pipeline_state.step_states[step_id]
return not step_state.invalid_details
can_be_skipped(self, step_id)
¶Check whether the processing of a step can be skipped.
Source code in kiara/models/module/pipeline/controller.py
def can_be_skipped(self, step_id: str) -> bool:
"""Check whether the processing of a step can be skipped."""
required = self.pipeline.structure.step_is_required(step_id=step_id)
if required:
required = self.can_be_processed(step_id)
return required
current_pipeline_state(self)
¶Source code in kiara/models/module/pipeline/controller.py
def current_pipeline_state(self) -> PipelineDetails:
if self._pipeline_details is None:
self._pipeline_details = self.pipeline.get_pipeline_details()
return self._pipeline_details
pipeline_is_ready(self)
¶Return whether the pipeline is ready to be processed.
A True result means that all pipeline inputs are set with valid values, and therefore every step within the
pipeline can be processed.
Returns:
| Type | Description |
|---|---|
bool |
whether the pipeline can be processed as a whole ( |
Source code in kiara/models/module/pipeline/controller.py
def pipeline_is_ready(self) -> bool:
"""Return whether the pipeline is ready to be processed.
A ``True`` result means that all pipeline inputs are set with valid values, and therefore every step within the
pipeline can be processed.
Returns:
whether the pipeline can be processed as a whole (``True``) or not (``False``)
"""
pipeline_inputs = self.pipeline._all_values.get_alias("pipeline.inputs")
assert pipeline_inputs is not None
return pipeline_inputs.all_items_valid
process_step(self, step_id, wait=False)
¶Kick off processing for the step with the provided id.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
step_id |
str |
the id of the step that should be started |
required |
Source code in kiara/models/module/pipeline/controller.py
def process_step(self, step_id: str, wait: bool = False) -> uuid.UUID:
"""Kick off processing for the step with the provided id.
Arguments:
step_id: the id of the step that should be started
"""
job_config = self.pipeline.create_job_config_for_step(step_id)
job_id = self._job_registry.execute_job(job_config=job_config)
# job_id = self._processor.create_job(job_config=job_config)
# self._processor.queue_job(job_id=job_id)
if wait:
self._job_registry.wait_for(job_id)
return job_id
set_processing_results(self, job_ids)
¶Source code in kiara/models/module/pipeline/controller.py
def set_processing_results(self, job_ids: Mapping[str, uuid.UUID]):
self._job_registry.wait_for(*job_ids.values())
combined_outputs = {}
for step_id, job_id in job_ids.items():
record = self._job_registry.get_job_record_in_session(job_id=job_id)
combined_outputs[step_id] = record.outputs
self.pipeline.set_multiple_step_outputs(
changed_outputs=combined_outputs, notify_listeners=True
)
pipeline
¶
Pipeline
¶An instance of a [PipelineStructure][kiara.pipeline.structure.PipelineStructure] that holds state for all of the inputs/outputs of the steps within.
Source code in kiara/models/module/pipeline/pipeline.py
class Pipeline(object):
"""An instance of a [PipelineStructure][kiara.pipeline.structure.PipelineStructure] that holds state for all of the inputs/outputs of the steps within."""
def __init__(self, structure: PipelineStructure, data_registry: DataRegistry):
self._id: uuid.UUID = uuid.uuid4()
self._structure: PipelineStructure = structure
self._value_refs: Mapping[AliasValueMap, Iterable[ValueRef]] = None # type: ignore
# self._status: StepStatus = StepStatus.STALE
self._steps_by_stage: Dict[int, Dict[str, PipelineStep]] = None # type: ignore
self._inputs_by_stage: Dict[int, List[str]] = None # type: ignore
self._outputs_by_stage: Dict[int, List[str]] = None # type: ignore
self._data_registry: DataRegistry = data_registry
self._all_values: AliasValueMap = None # type: ignore
self._init_values()
self._listeners: List[PipelineListener] = []
# self._update_status()
@property
def pipeline_id(self) -> uuid.UUID:
return self._id
@property
def kiara_id(self) -> uuid.UUID:
return self._data_registry.kiara_id
def _init_values(self):
"""Initialize this object. This should only be called once.
Basically, this goes through all the inputs and outputs of all steps, and 'allocates' a PipelineValueInfo object
for each of them. In case where output/input or pipeline-input/input points are connected, only one
value item is allocated, since those refer to the same value.
"""
values = AliasValueMap(
alias=str(self.id), version=0, assoc_value=None, values_schema={}
)
values._data_registry = self._data_registry
for field_name, schema in self._structure.pipeline_inputs_schema.items():
values.set_alias_schema(f"pipeline.inputs.{field_name}", schema=schema)
for field_name, schema in self._structure.pipeline_outputs_schema.items():
values.set_alias_schema(f"pipeline.outputs.{field_name}", schema=schema)
for step_id in self.step_ids:
step = self.get_step(step_id)
for field_name, value_schema in step.module.inputs_schema.items():
values.set_alias_schema(
f"steps.{step_id}.inputs.{field_name}", schema=value_schema
)
for field_name, value_schema in step.module.outputs_schema.items():
values.set_alias_schema(
f"steps.{step_id}.outputs.{field_name}", schema=value_schema
)
self._all_values = values
def __eq__(self, other):
if not isinstance(other, Pipeline):
return False
return self._id == other._id
def __hash__(self):
return hash(self._id)
def add_listener(self, listener: PipelineListener):
self._listeners.append(listener)
@property
def id(self) -> uuid.UUID:
return self._id
@property
def structure(self) -> PipelineStructure:
return self._structure
def get_current_pipeline_inputs(self) -> Dict[str, uuid.UUID]:
"""All (pipeline) input values of this pipeline."""
alias_map = self._all_values.get_alias("pipeline.inputs")
return alias_map.get_all_value_ids() # type: ignore
def get_current_pipeline_outputs(self) -> Dict[str, uuid.UUID]:
"""All (pipeline) output values of this pipeline."""
alias_map = self._all_values.get_alias("pipeline.outputs")
return alias_map.get_all_value_ids() # type: ignore
def get_current_step_inputs(self, step_id) -> Dict[str, uuid.UUID]:
alias_map = self._all_values.get_alias(f"steps.{step_id}.inputs")
return alias_map.get_all_value_ids() # type: ignore
def get_current_step_outputs(self, step_id) -> Dict[str, uuid.UUID]:
alias_map = self._all_values.get_alias(f"steps.{step_id}.outputs")
return alias_map.get_all_value_ids() # type: ignore
def get_inputs_for_steps(self, *step_ids: str) -> Dict[str, Dict[str, uuid.UUID]]:
"""Retrieve value ids for the inputs of the specified steps (or all steps, if no argument provided."""
result = {}
for step_id in self._structure.step_ids:
if step_ids and step_id not in step_ids:
continue
ids = self.get_current_step_inputs(step_id=step_id)
result[step_id] = ids
return result
def get_outputs_for_steps(self, *step_ids: str) -> Dict[str, Dict[str, uuid.UUID]]:
"""Retrieve value ids for the outputs of the specified steps (or all steps, if no argument provided."""
result = {}
for step_id in self._structure.step_ids:
if step_ids and step_id not in step_ids:
continue
ids = self.get_current_step_outputs(step_id=step_id)
result[step_id] = ids
return result
def _notify_pipeline_listeners(self, event: PipelineEvent):
for listener in self._listeners:
listener._pipeline_event_occurred(event=event)
def get_pipeline_details(self) -> PipelineDetails:
pipeline_inputs = self._all_values.get_alias("pipeline.inputs")
pipeline_outputs = self._all_values.get_alias("pipeline.outputs")
assert pipeline_inputs is not None
assert pipeline_outputs is not None
invalid = pipeline_inputs.check_invalid()
if not invalid:
status = StepStatus.INPUTS_READY
step_outputs = self._all_values.get_alias("pipeline.outputs")
assert step_outputs is not None
invalid_outputs = step_outputs.check_invalid()
# TODO: also check that all the pedigrees match up with current inputs
if not invalid_outputs:
status = StepStatus.RESULTS_READY
else:
status = StepStatus.INPUTS_INVALID
step_states = {}
for step_id in self._structure.step_ids:
d = self.get_step_details(step_id)
step_states[step_id] = d
details = PipelineDetails.construct(
kiara_id=self._data_registry.kiara_id,
pipeline_id=self.pipeline_id,
pipeline_status=status,
pipeline_inputs=pipeline_inputs.get_all_value_ids(),
pipeline_outputs=pipeline_outputs.get_all_value_ids(),
invalid_details=invalid,
step_states=step_states,
)
return details
def get_step_details(self, step_id: str) -> StepDetails:
step_input_ids = self.get_current_step_inputs(step_id=step_id)
step_output_ids = self.get_current_step_outputs(step_id=step_id)
step_inputs = self._all_values.get_alias(f"steps.{step_id}.inputs")
assert step_inputs is not None
invalid = step_inputs.check_invalid()
processing_stage = self._structure.get_processing_stage(step_id)
if not invalid:
status = StepStatus.INPUTS_READY
step_outputs = self._all_values.get_alias(f"steps.{step_id}.outputs")
assert step_outputs is not None
invalid_outputs = step_outputs.check_invalid()
# TODO: also check that all the pedigrees match up with current inputs
if not invalid_outputs:
status = StepStatus.RESULTS_READY
else:
status = StepStatus.INPUTS_INVALID
details = StepDetails.construct(
kiara_id=self._data_registry.kiara_id,
pipeline_id=self.pipeline_id,
step_id=step_id,
status=status,
inputs=step_input_ids,
outputs=step_output_ids,
invalid_details=invalid,
processing_stage=processing_stage,
)
return details
def set_pipeline_inputs(
self,
inputs: Mapping[str, Any],
sync_to_step_inputs: bool = True,
notify_listeners: bool = True,
) -> Mapping[str, Mapping[str, Mapping[str, ChangedValue]]]:
values_to_set: Dict[str, uuid.UUID] = {}
for k, v in inputs.items():
if v is None:
values_to_set[k] = NONE_VALUE_ID
else:
alias_map = self._all_values.get_alias("pipeline.inputs")
assert alias_map is not None
# dbg(alias_map.__dict__)
schema = alias_map.values_schema[k]
value = self._data_registry.register_data(
data=v, schema=schema, pedigree=ORPHAN, reuse_existing=True
)
values_to_set[k] = value.value_id
changed_pipeline_inputs = self._set_values("pipeline.inputs", **values_to_set)
changed_results = {"__pipeline__": {"inputs": changed_pipeline_inputs}}
if sync_to_step_inputs:
changed = self.sync_pipeline_inputs(notify_listeners=False)
dpath.util.merge(changed_results, changed) # type: ignore
if notify_listeners:
event = PipelineEvent.create_event(pipeline=self, changed=changed_results)
self._notify_pipeline_listeners(event)
return changed_results
def sync_pipeline_inputs(
self, notify_listeners: bool = True
) -> Mapping[str, Mapping[str, Mapping[str, ChangedValue]]]:
pipeline_inputs = self.get_current_pipeline_inputs()
values_to_sync: Dict[str, Dict[str, Optional[uuid.UUID]]] = {}
for field_name, ref in self._structure.pipeline_input_refs.items():
for step_input in ref.connected_inputs:
step_inputs = self.get_current_step_inputs(step_input.step_id)
if step_inputs[step_input.value_name] != pipeline_inputs[field_name]:
values_to_sync.setdefault(step_input.step_id, {})[
step_input.value_name
] = pipeline_inputs[field_name]
results: Dict[str, Mapping[str, Mapping[str, ChangedValue]]] = {}
for step_id in values_to_sync.keys():
values = values_to_sync[step_id]
step_changed = self._set_step_inputs(step_id=step_id, inputs=values)
dpath.util.merge(results, step_changed) # type: ignore
if notify_listeners:
event = PipelineEvent.create_event(pipeline=self, changed=results)
self._notify_pipeline_listeners(event)
return results
def _set_step_inputs(
self, step_id: str, inputs: Mapping[str, Optional[uuid.UUID]]
) -> Mapping[str, Mapping[str, Mapping[str, ChangedValue]]]:
changed_step_inputs = self._set_values(f"steps.{step_id}.inputs", **inputs)
if not changed_step_inputs:
return {}
result: Dict[str, Dict[str, Dict[str, ChangedValue]]] = {
step_id: {"inputs": changed_step_inputs}
}
step_outputs = self._structure.get_step_output_refs(step_id=step_id)
null_outputs = {k: None for k in step_outputs.keys()}
changed_outputs = self.set_step_outputs(
step_id=step_id, outputs=null_outputs, notify_listeners=False
)
assert step_id not in changed_outputs.keys()
result.update(changed_outputs) # type: ignore
return result
def set_multiple_step_outputs(
self,
changed_outputs: Mapping[str, Mapping[str, Optional[uuid.UUID]]],
notify_listeners: bool = True,
) -> Mapping[str, Mapping[str, Mapping[str, ChangedValue]]]:
results: Dict[str, Dict[str, Dict[str, ChangedValue]]] = {}
for step_id, outputs in changed_outputs.items():
step_results = self.set_step_outputs(
step_id=step_id, outputs=outputs, notify_listeners=False
)
dpath.util.merge(results, step_results) # type: ignore
if notify_listeners:
event = PipelineEvent.create_event(pipeline=self, changed=results)
self._notify_pipeline_listeners(event)
return results
def set_step_outputs(
self,
step_id: str,
outputs: Mapping[str, Optional[uuid.UUID]],
notify_listeners: bool = True,
) -> Mapping[str, Mapping[str, Mapping[str, ChangedValue]]]:
# make sure pedigrees match with respective inputs?
changed_step_outputs = self._set_values(f"steps.{step_id}.outputs", **outputs)
if not changed_step_outputs:
return {}
result: Dict[str, Dict[str, Dict[str, ChangedValue]]] = {
step_id: {"outputs": changed_step_outputs}
}
output_refs = self._structure.get_step_output_refs(step_id=step_id)
pipeline_outputs: Dict[str, Optional[uuid.UUID]] = {}
inputs_to_set: Dict[str, Dict[str, Optional[uuid.UUID]]] = {}
for field_name, ref in output_refs.items():
if ref.pipeline_output:
assert ref.pipeline_output not in pipeline_outputs.keys()
pipeline_outputs[ref.pipeline_output] = outputs[field_name]
for input_ref in ref.connected_inputs:
inputs_to_set.setdefault(input_ref.step_id, {})[
input_ref.value_name
] = outputs[field_name]
for step_id, step_inputs in inputs_to_set.items():
changed_step_fields = self._set_step_inputs(
step_id=step_id, inputs=step_inputs
)
dpath.util.merge(result, changed_step_fields) # type: ignore
if pipeline_outputs:
changed_pipeline_outputs = self._set_pipeline_outputs(**pipeline_outputs)
dpath.util.merge( # type: ignore
result, {"__pipeline__": {"outputs": changed_pipeline_outputs}}
)
if notify_listeners:
event = PipelineEvent.create_event(pipeline=self, changed=result)
self._notify_pipeline_listeners(event)
return result
def _set_pipeline_outputs(
self, **outputs: Optional[uuid.UUID]
) -> Mapping[str, ChangedValue]:
changed_pipeline_outputs = self._set_values("pipeline.outputs", **outputs)
return changed_pipeline_outputs
def _set_values(
self, alias: str, **values: Optional[uuid.UUID]
) -> Dict[str, ChangedValue]:
"""Set values (value-ids) for the sub-alias-map with the specified alias path."""
invalid = {}
for k in values.keys():
_alias = self._all_values.get_alias(alias)
assert _alias is not None
if k not in _alias.values_schema.keys():
invalid[
k
] = f"Invalid field '{k}'. Available fields: {', '.join(self.get_current_pipeline_inputs().keys())}"
if invalid:
raise InvalidValuesException(invalid_values=invalid)
alias_map: Optional[AliasValueMap] = self._all_values.get_alias(alias)
assert alias_map is not None
values_to_set: Dict[str, Optional[uuid.UUID]] = {}
current: Dict[str, Optional[uuid.UUID]] = {}
changed: Dict[str, ChangedValue] = {}
for field_name, new_value in values.items():
current_value = self._all_values.get_alias(f"{alias}.{field_name}")
if current_value is not None:
current_value_id = current_value.assoc_value
else:
current_value_id = None
current[field_name] = current_value_id
if current_value_id != new_value:
values_to_set[field_name] = new_value
changed[field_name] = ChangedValue(old=current_value_id, new=new_value)
_alias = self._all_values.get_alias(alias)
assert _alias is not None
_alias.set_aliases(**values_to_set)
return changed
@property
def step_ids(self) -> Iterable[str]:
"""Return all ids of the steps of this pipeline."""
return self._structure.step_ids
def get_step(self, step_id: str) -> PipelineStep:
"""Return the object representing a step in this workflow, identified by the step id."""
return self._structure.get_step(step_id)
def get_steps_by_stage(
self,
) -> Mapping[int, Mapping[str, PipelineStep]]:
"""Return a all pipeline steps, ordered by stage they belong to."""
if self._steps_by_stage is not None:
return self._steps_by_stage
result: Dict[int, Dict[str, PipelineStep]] = {}
for step_id in self.step_ids:
step = self.get_step(step_id)
stage = self._structure.get_processing_stage(step.step_id)
assert stage is not None
result.setdefault(stage, {})[step_id] = step
self._steps_by_stage = result
return self._steps_by_stage
def create_job_config_for_step(self, step_id: str) -> JobConfig:
step_inputs: Mapping[str, Optional[uuid.UUID]] = self.get_current_step_inputs(
step_id
)
step_details: StepDetails = self.get_step_details(step_id=step_id)
step: PipelineStep = self.get_step(step_id=step_id)
# if the inputs are not valid, ignore this step
if step_details.status == StepStatus.INPUTS_INVALID:
invalid_details = step_details.invalid_details
assert invalid_details is not None
msg = f"Can't execute step '{step_id}', invalid inputs: {', '.join(invalid_details.keys())}"
raise InvalidValuesException(msg=msg, invalid_values=invalid_details)
job_config = JobConfig.create_from_module(
data_registry=self._data_registry, module=step.module, inputs=step_inputs
)
return job_config
id: UUID
property
readonly
¶kiara_id: UUID
property
readonly
¶pipeline_id: UUID
property
readonly
¶step_ids: Iterable[str]
property
readonly
¶Return all ids of the steps of this pipeline.
structure: PipelineStructure
property
readonly
¶add_listener(self, listener)
¶Source code in kiara/models/module/pipeline/pipeline.py
def add_listener(self, listener: PipelineListener):
self._listeners.append(listener)
create_job_config_for_step(self, step_id)
¶Source code in kiara/models/module/pipeline/pipeline.py
def create_job_config_for_step(self, step_id: str) -> JobConfig:
step_inputs: Mapping[str, Optional[uuid.UUID]] = self.get_current_step_inputs(
step_id
)
step_details: StepDetails = self.get_step_details(step_id=step_id)
step: PipelineStep = self.get_step(step_id=step_id)
# if the inputs are not valid, ignore this step
if step_details.status == StepStatus.INPUTS_INVALID:
invalid_details = step_details.invalid_details
assert invalid_details is not None
msg = f"Can't execute step '{step_id}', invalid inputs: {', '.join(invalid_details.keys())}"
raise InvalidValuesException(msg=msg, invalid_values=invalid_details)
job_config = JobConfig.create_from_module(
data_registry=self._data_registry, module=step.module, inputs=step_inputs
)
return job_config
get_current_pipeline_inputs(self)
¶All (pipeline) input values of this pipeline.
Source code in kiara/models/module/pipeline/pipeline.py
def get_current_pipeline_inputs(self) -> Dict[str, uuid.UUID]:
"""All (pipeline) input values of this pipeline."""
alias_map = self._all_values.get_alias("pipeline.inputs")
return alias_map.get_all_value_ids() # type: ignore
get_current_pipeline_outputs(self)
¶All (pipeline) output values of this pipeline.
Source code in kiara/models/module/pipeline/pipeline.py
def get_current_pipeline_outputs(self) -> Dict[str, uuid.UUID]:
"""All (pipeline) output values of this pipeline."""
alias_map = self._all_values.get_alias("pipeline.outputs")
return alias_map.get_all_value_ids() # type: ignore
get_current_step_inputs(self, step_id)
¶Source code in kiara/models/module/pipeline/pipeline.py
def get_current_step_inputs(self, step_id) -> Dict[str, uuid.UUID]:
alias_map = self._all_values.get_alias(f"steps.{step_id}.inputs")
return alias_map.get_all_value_ids() # type: ignore
get_current_step_outputs(self, step_id)
¶Source code in kiara/models/module/pipeline/pipeline.py
def get_current_step_outputs(self, step_id) -> Dict[str, uuid.UUID]:
alias_map = self._all_values.get_alias(f"steps.{step_id}.outputs")
return alias_map.get_all_value_ids() # type: ignore
get_inputs_for_steps(self, *step_ids)
¶Retrieve value ids for the inputs of the specified steps (or all steps, if no argument provided.
Source code in kiara/models/module/pipeline/pipeline.py
def get_inputs_for_steps(self, *step_ids: str) -> Dict[str, Dict[str, uuid.UUID]]:
"""Retrieve value ids for the inputs of the specified steps (or all steps, if no argument provided."""
result = {}
for step_id in self._structure.step_ids:
if step_ids and step_id not in step_ids:
continue
ids = self.get_current_step_inputs(step_id=step_id)
result[step_id] = ids
return result
get_outputs_for_steps(self, *step_ids)
¶Retrieve value ids for the outputs of the specified steps (or all steps, if no argument provided.
Source code in kiara/models/module/pipeline/pipeline.py
def get_outputs_for_steps(self, *step_ids: str) -> Dict[str, Dict[str, uuid.UUID]]:
"""Retrieve value ids for the outputs of the specified steps (or all steps, if no argument provided."""
result = {}
for step_id in self._structure.step_ids:
if step_ids and step_id not in step_ids:
continue
ids = self.get_current_step_outputs(step_id=step_id)
result[step_id] = ids
return result
get_pipeline_details(self)
¶Source code in kiara/models/module/pipeline/pipeline.py
def get_pipeline_details(self) -> PipelineDetails:
pipeline_inputs = self._all_values.get_alias("pipeline.inputs")
pipeline_outputs = self._all_values.get_alias("pipeline.outputs")
assert pipeline_inputs is not None
assert pipeline_outputs is not None
invalid = pipeline_inputs.check_invalid()
if not invalid:
status = StepStatus.INPUTS_READY
step_outputs = self._all_values.get_alias("pipeline.outputs")
assert step_outputs is not None
invalid_outputs = step_outputs.check_invalid()
# TODO: also check that all the pedigrees match up with current inputs
if not invalid_outputs:
status = StepStatus.RESULTS_READY
else:
status = StepStatus.INPUTS_INVALID
step_states = {}
for step_id in self._structure.step_ids:
d = self.get_step_details(step_id)
step_states[step_id] = d
details = PipelineDetails.construct(
kiara_id=self._data_registry.kiara_id,
pipeline_id=self.pipeline_id,
pipeline_status=status,
pipeline_inputs=pipeline_inputs.get_all_value_ids(),
pipeline_outputs=pipeline_outputs.get_all_value_ids(),
invalid_details=invalid,
step_states=step_states,
)
return details
get_step(self, step_id)
¶Return the object representing a step in this workflow, identified by the step id.
Source code in kiara/models/module/pipeline/pipeline.py
def get_step(self, step_id: str) -> PipelineStep:
"""Return the object representing a step in this workflow, identified by the step id."""
return self._structure.get_step(step_id)
get_step_details(self, step_id)
¶Source code in kiara/models/module/pipeline/pipeline.py
def get_step_details(self, step_id: str) -> StepDetails:
step_input_ids = self.get_current_step_inputs(step_id=step_id)
step_output_ids = self.get_current_step_outputs(step_id=step_id)
step_inputs = self._all_values.get_alias(f"steps.{step_id}.inputs")
assert step_inputs is not None
invalid = step_inputs.check_invalid()
processing_stage = self._structure.get_processing_stage(step_id)
if not invalid:
status = StepStatus.INPUTS_READY
step_outputs = self._all_values.get_alias(f"steps.{step_id}.outputs")
assert step_outputs is not None
invalid_outputs = step_outputs.check_invalid()
# TODO: also check that all the pedigrees match up with current inputs
if not invalid_outputs:
status = StepStatus.RESULTS_READY
else:
status = StepStatus.INPUTS_INVALID
details = StepDetails.construct(
kiara_id=self._data_registry.kiara_id,
pipeline_id=self.pipeline_id,
step_id=step_id,
status=status,
inputs=step_input_ids,
outputs=step_output_ids,
invalid_details=invalid,
processing_stage=processing_stage,
)
return details
get_steps_by_stage(self)
¶Return a all pipeline steps, ordered by stage they belong to.
Source code in kiara/models/module/pipeline/pipeline.py
def get_steps_by_stage(
self,
) -> Mapping[int, Mapping[str, PipelineStep]]:
"""Return a all pipeline steps, ordered by stage they belong to."""
if self._steps_by_stage is not None:
return self._steps_by_stage
result: Dict[int, Dict[str, PipelineStep]] = {}
for step_id in self.step_ids:
step = self.get_step(step_id)
stage = self._structure.get_processing_stage(step.step_id)
assert stage is not None
result.setdefault(stage, {})[step_id] = step
self._steps_by_stage = result
return self._steps_by_stage
set_multiple_step_outputs(self, changed_outputs, notify_listeners=True)
¶Source code in kiara/models/module/pipeline/pipeline.py
def set_multiple_step_outputs(
self,
changed_outputs: Mapping[str, Mapping[str, Optional[uuid.UUID]]],
notify_listeners: bool = True,
) -> Mapping[str, Mapping[str, Mapping[str, ChangedValue]]]:
results: Dict[str, Dict[str, Dict[str, ChangedValue]]] = {}
for step_id, outputs in changed_outputs.items():
step_results = self.set_step_outputs(
step_id=step_id, outputs=outputs, notify_listeners=False
)
dpath.util.merge(results, step_results) # type: ignore
if notify_listeners:
event = PipelineEvent.create_event(pipeline=self, changed=results)
self._notify_pipeline_listeners(event)
return results
set_pipeline_inputs(self, inputs, sync_to_step_inputs=True, notify_listeners=True)
¶Source code in kiara/models/module/pipeline/pipeline.py
def set_pipeline_inputs(
self,
inputs: Mapping[str, Any],
sync_to_step_inputs: bool = True,
notify_listeners: bool = True,
) -> Mapping[str, Mapping[str, Mapping[str, ChangedValue]]]:
values_to_set: Dict[str, uuid.UUID] = {}
for k, v in inputs.items():
if v is None:
values_to_set[k] = NONE_VALUE_ID
else:
alias_map = self._all_values.get_alias("pipeline.inputs")
assert alias_map is not None
# dbg(alias_map.__dict__)
schema = alias_map.values_schema[k]
value = self._data_registry.register_data(
data=v, schema=schema, pedigree=ORPHAN, reuse_existing=True
)
values_to_set[k] = value.value_id
changed_pipeline_inputs = self._set_values("pipeline.inputs", **values_to_set)
changed_results = {"__pipeline__": {"inputs": changed_pipeline_inputs}}
if sync_to_step_inputs:
changed = self.sync_pipeline_inputs(notify_listeners=False)
dpath.util.merge(changed_results, changed) # type: ignore
if notify_listeners:
event = PipelineEvent.create_event(pipeline=self, changed=changed_results)
self._notify_pipeline_listeners(event)
return changed_results
set_step_outputs(self, step_id, outputs, notify_listeners=True)
¶Source code in kiara/models/module/pipeline/pipeline.py
def set_step_outputs(
self,
step_id: str,
outputs: Mapping[str, Optional[uuid.UUID]],
notify_listeners: bool = True,
) -> Mapping[str, Mapping[str, Mapping[str, ChangedValue]]]:
# make sure pedigrees match with respective inputs?
changed_step_outputs = self._set_values(f"steps.{step_id}.outputs", **outputs)
if not changed_step_outputs:
return {}
result: Dict[str, Dict[str, Dict[str, ChangedValue]]] = {
step_id: {"outputs": changed_step_outputs}
}
output_refs = self._structure.get_step_output_refs(step_id=step_id)
pipeline_outputs: Dict[str, Optional[uuid.UUID]] = {}
inputs_to_set: Dict[str, Dict[str, Optional[uuid.UUID]]] = {}
for field_name, ref in output_refs.items():
if ref.pipeline_output:
assert ref.pipeline_output not in pipeline_outputs.keys()
pipeline_outputs[ref.pipeline_output] = outputs[field_name]
for input_ref in ref.connected_inputs:
inputs_to_set.setdefault(input_ref.step_id, {})[
input_ref.value_name
] = outputs[field_name]
for step_id, step_inputs in inputs_to_set.items():
changed_step_fields = self._set_step_inputs(
step_id=step_id, inputs=step_inputs
)
dpath.util.merge(result, changed_step_fields) # type: ignore
if pipeline_outputs:
changed_pipeline_outputs = self._set_pipeline_outputs(**pipeline_outputs)
dpath.util.merge( # type: ignore
result, {"__pipeline__": {"outputs": changed_pipeline_outputs}}
)
if notify_listeners:
event = PipelineEvent.create_event(pipeline=self, changed=result)
self._notify_pipeline_listeners(event)
return result
sync_pipeline_inputs(self, notify_listeners=True)
¶Source code in kiara/models/module/pipeline/pipeline.py
def sync_pipeline_inputs(
self, notify_listeners: bool = True
) -> Mapping[str, Mapping[str, Mapping[str, ChangedValue]]]:
pipeline_inputs = self.get_current_pipeline_inputs()
values_to_sync: Dict[str, Dict[str, Optional[uuid.UUID]]] = {}
for field_name, ref in self._structure.pipeline_input_refs.items():
for step_input in ref.connected_inputs:
step_inputs = self.get_current_step_inputs(step_input.step_id)
if step_inputs[step_input.value_name] != pipeline_inputs[field_name]:
values_to_sync.setdefault(step_input.step_id, {})[
step_input.value_name
] = pipeline_inputs[field_name]
results: Dict[str, Mapping[str, Mapping[str, ChangedValue]]] = {}
for step_id in values_to_sync.keys():
values = values_to_sync[step_id]
step_changed = self._set_step_inputs(step_id=step_id, inputs=values)
dpath.util.merge(results, step_changed) # type: ignore
if notify_listeners:
event = PipelineEvent.create_event(pipeline=self, changed=results)
self._notify_pipeline_listeners(event)
return results
PipelineListener (ABC)
¶Source code in kiara/models/module/pipeline/pipeline.py
class PipelineListener(abc.ABC):
@abc.abstractmethod
def _pipeline_event_occurred(self, event: PipelineEvent):
pass
structure
¶
PipelineStructure (KiaraModel)
pydantic-model
¶An object that holds one or several steps, and describes the connections between them.
Source code in kiara/models/module/pipeline/structure.py
class PipelineStructure(KiaraModel):
"""An object that holds one or several steps, and describes the connections between them."""
pipeline_config: PipelineConfig = Field(
description="The underlying pipeline config."
)
steps: List[PipelineStep] = Field(description="The pipeline steps ")
input_aliases: Dict[str, str] = Field(description="The (resolved) input aliases.")
output_aliases: Dict[str, str] = Field(description="The (resolved) output aliases.")
@root_validator(pre=True)
def validate_pipeline_config(cls, values):
pipeline_config = values.get("pipeline_config", None)
if not pipeline_config:
raise ValueError("No 'pipeline_config' provided.")
if len(values) != 1:
raise ValueError(
"Only 'pipeline_config' key allowed when creating a pipeline structure object."
)
_config: PipelineConfig = pipeline_config
_steps: List[PipelineStep] = list(_config.steps)
_input_aliases: Dict[str, str] = dict(_config.input_aliases)
_output_aliases: Dict[str, str] = dict(_config.output_aliases)
values["steps"] = _steps
values["input_aliases"] = _input_aliases
values["output_aliases"] = _output_aliases
return values
# this is hardcoded for now
_add_all_workflow_outputs: bool = PrivateAttr(default=False)
_constants: Dict[str, Any] = PrivateAttr(default=None) # type: ignore
_defaults: Dict[str, Any] = PrivateAttr(None) # type: ignore
_execution_graph: nx.DiGraph = PrivateAttr(None) # type: ignore
_data_flow_graph: nx.DiGraph = PrivateAttr(None) # type: ignore
_data_flow_graph_simple: nx.DiGraph = PrivateAttr(None) # type: ignore
_processing_stages: List[List[str]] = PrivateAttr(None) # type: ignore
# holds details about the (current) processing steps contained in this workflow
_steps_details: Dict[str, Any] = PrivateAttr(None) # type: ignore
def _retrieve_data_to_hash(self) -> Any:
return {
"steps": [step.model_data_hash for step in self.steps],
"input_aliases": self.input_aliases,
"output_aliases": self.output_aliases,
}
def _retrieve_id(self) -> str:
return self.pipeline_config.model_id
def _retrieve_category_id(self) -> str:
return PIPELINE_STRUCTURE_TYPE_CATEGORY_ID
@property
def steps_details(self) -> Mapping[str, Any]:
if self._steps_details is None:
self._process_steps()
return self._steps_details # type: ignore
@property
def step_ids(self) -> Iterable[str]:
if self._steps_details is None:
self._process_steps()
return self._steps_details.keys() # type: ignore
@property
def constants(self) -> Mapping[str, Any]:
if self._constants is None:
self._process_steps()
return self._constants # type: ignore
@property
def defaults(self) -> Mapping[str, Any]:
if self._defaults is None:
self._process_steps()
return self._defaults # type: ignore
def get_step(self, step_id: str) -> PipelineStep:
d = self.steps_details.get(step_id, None)
if d is None:
raise Exception(f"No module with id: {step_id}")
return d["step"]
def get_step_input_refs(self, step_id: str) -> Mapping[str, StepInputRef]:
d = self.steps_details.get(step_id, None)
if d is None:
raise Exception(f"No module with id: {step_id}")
return d["inputs"]
def get_step_output_refs(self, step_id: str) -> Mapping[str, StepOutputRef]:
d = self.steps_details.get(step_id, None)
if d is None:
raise Exception(f"No module with id: {step_id}")
return d["outputs"]
def get_step_details(self, step_id: str) -> Mapping[str, Any]:
d = self.steps_details.get(step_id, None)
if d is None:
raise Exception(f"No module with id: {step_id}")
return d
@property
def execution_graph(self) -> nx.DiGraph:
if self._execution_graph is None:
self._process_steps()
return self._execution_graph
@property
def data_flow_graph(self) -> nx.DiGraph:
if self._data_flow_graph is None:
self._process_steps()
return self._data_flow_graph
@property
def data_flow_graph_simple(self) -> nx.DiGraph:
if self._data_flow_graph_simple is None:
self._process_steps()
return self._data_flow_graph_simple
@property
def processing_stages(self) -> List[List[str]]:
if self._steps_details is None:
self._process_steps()
return self._processing_stages
@lru_cache()
def _get_node_of_type(self, node_type: str):
if self._steps_details is None:
self._process_steps()
return [
node
for node, attr in self._data_flow_graph.nodes(data=True)
if attr["type"] == node_type
]
@property
def steps_input_refs(self) -> Dict[str, StepInputRef]:
return {
node.alias: node
for node in self._get_node_of_type(node_type=StepInputRef.__name__)
}
@property
def steps_output_refs(self) -> Dict[str, StepOutputRef]:
return {
node.alias: node
for node in self._get_node_of_type(node_type=StepOutputRef.__name__)
}
@property
def pipeline_input_refs(self) -> Dict[str, PipelineInputRef]:
return {
node.value_name: node
for node in self._get_node_of_type(node_type=PipelineInputRef.__name__)
}
@property
def pipeline_output_refs(self) -> Dict[str, PipelineOutputRef]:
return {
node.value_name: node
for node in self._get_node_of_type(node_type=PipelineOutputRef.__name__)
}
@property
def pipeline_inputs_schema(self) -> Mapping[str, ValueSchema]:
return {
input_name: w_in.value_schema
for input_name, w_in in self.pipeline_input_refs.items()
}
@property
def pipeline_outputs_schema(self) -> Mapping[str, ValueSchema]:
return {
output_name: w_out.value_schema
for output_name, w_out in self.pipeline_output_refs.items()
}
def get_processing_stage(self, step_id: str) -> int:
"""Return the processing stage for the specified step_id.
Returns the stage nr (starting with '1').
"""
for index, stage in enumerate(self.processing_stages, start=1):
if step_id in stage:
return index
raise Exception(f"Invalid step id '{step_id}'.")
def step_is_required(self, step_id: str) -> bool:
"""Check if the specified step is required, or can be omitted."""
return self.get_step_details(step_id=step_id)["required"]
def _process_steps(self):
"""The core method of this class, it connects all the processing modules, their inputs and outputs."""
steps_details: Dict[str, Any] = {}
execution_graph = nx.DiGraph()
execution_graph.add_node("__root__")
data_flow_graph = nx.DiGraph()
data_flow_graph_simple = nx.DiGraph()
processing_stages = []
constants = {}
structure_defaults = {}
# temp variable, to hold all outputs
outputs: Dict[str, StepOutputRef] = {}
# process all pipeline and step outputs first
_temp_steps_map: Dict[str, PipelineStep] = {}
pipeline_outputs: Dict[str, PipelineOutputRef] = {}
for step in self.steps:
_temp_steps_map[step.step_id] = step
if step.step_id in steps_details.keys():
raise Exception(
f"Can't process steps: duplicate step_id '{step.step_id}'"
)
steps_details[step.step_id] = {
"step": step,
"outputs": {},
"inputs": {},
"required": True,
}
data_flow_graph.add_node(step, type="step")
# go through all the module outputs, create points for them and connect them to pipeline outputs
for output_name, schema in step.module.outputs_schema.items():
step_output = StepOutputRef(
value_name=output_name,
value_schema=schema,
step_id=step.step_id,
)
steps_details[step.step_id]["outputs"][output_name] = step_output
step_alias = generate_step_alias(step.step_id, output_name)
outputs[step_alias] = step_output
# step_output_name = generate_pipeline_endpoint_name(
# step_id=step.step_id, value_name=output_name
# )
step_output_name = f"{step.step_id}.{output_name}"
if not self.output_aliases:
raise NotImplementedError()
if step_output_name in self.output_aliases.keys():
step_output_name = self.output_aliases[step_output_name]
else:
if not self._add_all_workflow_outputs:
# this output is not interesting for the workflow
step_output_name = None
if step_output_name:
step_output_address = StepValueAddress(
step_id=step.step_id, value_name=output_name
)
pipeline_output = PipelineOutputRef(
value_name=step_output_name,
connected_output=step_output_address,
value_schema=schema,
)
pipeline_outputs[step_output_name] = pipeline_output
step_output.pipeline_output = pipeline_output.value_name
data_flow_graph.add_node(
pipeline_output, type=PipelineOutputRef.__name__
)
data_flow_graph.add_edge(step_output, pipeline_output)
data_flow_graph_simple.add_node(
pipeline_output, type=PipelineOutputRef.__name__
)
data_flow_graph_simple.add_edge(step, pipeline_output)
data_flow_graph.add_node(step_output, type=StepOutputRef.__name__)
data_flow_graph.add_edge(step, step_output)
# now process inputs, and connect them to the appropriate output/pipeline-input points
existing_pipeline_input_points: Dict[str, PipelineInputRef] = {}
for step in self.steps:
other_step_dependency: Set = set()
# go through all the inputs of a module, create input points and connect them to either
# other module outputs, or pipeline inputs (which need to be created)
module_constants: Mapping[str, Any] = step.module.get_config_value(
"constants"
)
for input_name, schema in step.module.inputs_schema.items():
matching_input_links: List[StepValueAddress] = []
is_constant = input_name in module_constants.keys()
for value_name, input_links in step.input_links.items():
if value_name == input_name:
for input_link in input_links:
if input_link in matching_input_links:
raise Exception(f"Duplicate input link: {input_link}")
matching_input_links.append(input_link)
if matching_input_links:
# this means we connect to other steps output
connected_output_points: List[StepOutputRef] = []
connected_outputs: List[StepValueAddress] = []
for input_link in matching_input_links:
output_id = generate_step_alias(
input_link.step_id, input_link.value_name
)
if output_id not in outputs.keys():
raise Exception(
f"Can't connect input '{input_name}' for step '{step.step_id}': no output '{output_id}' available. Available output names: {', '.join(outputs.keys())}"
)
connected_output_points.append(outputs[output_id])
connected_outputs.append(input_link)
other_step_dependency.add(input_link.step_id)
step_input_point = StepInputRef(
step_id=step.step_id,
value_name=input_name,
value_schema=schema,
is_constant=is_constant,
connected_pipeline_input=None,
connected_outputs=connected_outputs,
)
for op in connected_output_points:
op.connected_inputs.append(step_input_point.address)
data_flow_graph.add_edge(op, step_input_point)
data_flow_graph_simple.add_edge(
_temp_steps_map[op.step_id], step_input_point
) # TODO: name edge
data_flow_graph_simple.add_edge(
step_input_point, step
) # TODO: name edge
else:
# this means we connect to pipeline input
# pipeline_input_name = generate_pipeline_endpoint_name(
# step_id=step.step_id, value_name=input_name
# )
pipeline_input_name = f"{step.step_id}.{input_name}"
# check whether this input has an alias associated with it
if not self.input_aliases:
raise NotImplementedError()
if pipeline_input_name in self.input_aliases.keys():
# this means we use the pipeline alias
pipeline_input_name = self.input_aliases[pipeline_input_name]
if pipeline_input_name in existing_pipeline_input_points.keys():
# we already created a pipeline input with this name
# TODO: check whether schema fits
connected_pipeline_input = existing_pipeline_input_points[
pipeline_input_name
]
assert connected_pipeline_input.is_constant == is_constant
else:
# we need to create the pipeline input
connected_pipeline_input = PipelineInputRef(
value_name=pipeline_input_name,
value_schema=schema,
is_constant=is_constant,
)
existing_pipeline_input_points[
pipeline_input_name
] = connected_pipeline_input
data_flow_graph.add_node(
connected_pipeline_input, type=PipelineInputRef.__name__
)
data_flow_graph_simple.add_node(
connected_pipeline_input, type=PipelineInputRef.__name__
)
if is_constant:
constants[
pipeline_input_name
] = step.module.get_config_value("constants")[input_name]
default_val = step.module.get_config_value("defaults").get(
input_name, None
)
if is_constant and default_val is not None:
raise Exception(
f"Module config invalid for step '{step.step_id}': both default value and constant provided for input '{input_name}'."
)
elif default_val is not None:
structure_defaults[pipeline_input_name] = default_val
step_input_point = StepInputRef(
step_id=step.step_id,
value_name=input_name,
value_schema=schema,
connected_pipeline_input=connected_pipeline_input.value_name,
connected_outputs=None,
)
connected_pipeline_input.connected_inputs.append(
step_input_point.address
)
data_flow_graph.add_edge(connected_pipeline_input, step_input_point)
data_flow_graph_simple.add_edge(connected_pipeline_input, step)
data_flow_graph.add_node(step_input_point, type=StepInputRef.__name__)
steps_details[step.step_id]["inputs"][input_name] = step_input_point
data_flow_graph.add_edge(step_input_point, step)
if other_step_dependency:
for module_id in other_step_dependency:
execution_graph.add_edge(module_id, step.step_id)
else:
execution_graph.add_edge("__root__", step.step_id)
# calculate execution order
path_lengths: Dict[str, int] = {}
for step in self.steps:
step_id = step.step_id
paths = list(nx.all_simple_paths(execution_graph, "__root__", step_id))
max_steps = max(paths, key=lambda x: len(x))
path_lengths[step_id] = len(max_steps) - 1
max_length = max(path_lengths.values())
for i in range(1, max_length + 1):
stage: List[str] = [m for m, length in path_lengths.items() if length == i]
processing_stages.append(stage)
for _step_id in stage:
steps_details[_step_id]["processing_stage"] = i
# steps_details[_step_id]["step"].processing_stage = i
self._constants = constants
self._defaults = structure_defaults
self._steps_details = steps_details
self._execution_graph = execution_graph
self._data_flow_graph = data_flow_graph
self._data_flow_graph_simple = data_flow_graph_simple
self._processing_stages = processing_stages
self._get_node_of_type.cache_clear()
# calculating which steps are always required to execute to compute one of the required pipeline outputs.
# this is done because in some cases it's possible that some steps can be skipped to execute if they
# don't have a valid input set, because the inputs downstream they are connecting to are 'non-required'
# optional_steps = []
last_stage = self._processing_stages[-1]
step_nodes: List[PipelineStep] = [
node
for node in self._data_flow_graph_simple.nodes
if isinstance(node, PipelineStep)
]
all_required_inputs = []
for step_id in last_stage:
step = self.get_step(step_id)
step_nodes.remove(step)
for k, s_inp in self.get_step_input_refs(step_id).items():
if not s_inp.value_schema.is_required():
continue
all_required_inputs.append(s_inp)
for pipeline_input in self.pipeline_input_refs.values():
for last_step_input in all_required_inputs:
try:
path = nx.shortest_path(
self._data_flow_graph_simple, pipeline_input, last_step_input
)
for p in path:
if p in step_nodes:
step_nodes.remove(p)
except (NetworkXNoPath, NodeNotFound):
pass
# print("NO PATH")
# print(f"{pipeline_input} -> {last_step_input}")
for s in step_nodes:
self._steps_details[s.step_id]["required"] = False
# s.required = False
for input_name, inp in self.pipeline_input_refs.items():
steps = set()
for ci in inp.connected_inputs:
steps.add(ci.step_id)
optional = True
for step_id in steps:
step = self.get_step(step_id)
if self._steps_details[step_id]["required"]:
optional = False
break
if optional:
inp.value_schema.optional = True
constants: Mapping[str, Any]
property
readonly
¶data_flow_graph: DiGraph
property
readonly
¶data_flow_graph_simple: DiGraph
property
readonly
¶defaults: Mapping[str, Any]
property
readonly
¶execution_graph: DiGraph
property
readonly
¶input_aliases: Dict[str, str]
pydantic-field
required
¶The (resolved) input aliases.
output_aliases: Dict[str, str]
pydantic-field
required
¶The (resolved) output aliases.
pipeline_config: PipelineConfig
pydantic-field
required
¶The underlying pipeline config.
pipeline_input_refs: Dict[str, kiara.models.module.pipeline.value_refs.PipelineInputRef]
property
readonly
¶pipeline_inputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
property
readonly
¶pipeline_output_refs: Dict[str, kiara.models.module.pipeline.value_refs.PipelineOutputRef]
property
readonly
¶pipeline_outputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
property
readonly
¶processing_stages: List[List[str]]
property
readonly
¶step_ids: Iterable[str]
property
readonly
¶steps: List[kiara.models.module.pipeline.PipelineStep]
pydantic-field
required
¶The pipeline steps
steps_details: Mapping[str, Any]
property
readonly
¶steps_input_refs: Dict[str, kiara.models.module.pipeline.value_refs.StepInputRef]
property
readonly
¶steps_output_refs: Dict[str, kiara.models.module.pipeline.value_refs.StepOutputRef]
property
readonly
¶get_processing_stage(self, step_id)
¶Return the processing stage for the specified step_id.
Returns the stage nr (starting with '1').
Source code in kiara/models/module/pipeline/structure.py
def get_processing_stage(self, step_id: str) -> int:
"""Return the processing stage for the specified step_id.
Returns the stage nr (starting with '1').
"""
for index, stage in enumerate(self.processing_stages, start=1):
if step_id in stage:
return index
raise Exception(f"Invalid step id '{step_id}'.")
get_step(self, step_id)
¶Source code in kiara/models/module/pipeline/structure.py
def get_step(self, step_id: str) -> PipelineStep:
d = self.steps_details.get(step_id, None)
if d is None:
raise Exception(f"No module with id: {step_id}")
return d["step"]
get_step_details(self, step_id)
¶Source code in kiara/models/module/pipeline/structure.py
def get_step_details(self, step_id: str) -> Mapping[str, Any]:
d = self.steps_details.get(step_id, None)
if d is None:
raise Exception(f"No module with id: {step_id}")
return d
get_step_input_refs(self, step_id)
¶Source code in kiara/models/module/pipeline/structure.py
def get_step_input_refs(self, step_id: str) -> Mapping[str, StepInputRef]:
d = self.steps_details.get(step_id, None)
if d is None:
raise Exception(f"No module with id: {step_id}")
return d["inputs"]
get_step_output_refs(self, step_id)
¶Source code in kiara/models/module/pipeline/structure.py
def get_step_output_refs(self, step_id: str) -> Mapping[str, StepOutputRef]:
d = self.steps_details.get(step_id, None)
if d is None:
raise Exception(f"No module with id: {step_id}")
return d["outputs"]
step_is_required(self, step_id)
¶Check if the specified step is required, or can be omitted.
Source code in kiara/models/module/pipeline/structure.py
def step_is_required(self, step_id: str) -> bool:
"""Check if the specified step is required, or can be omitted."""
return self.get_step_details(step_id=step_id)["required"]
validate_pipeline_config(values)
classmethod
¶Source code in kiara/models/module/pipeline/structure.py
@root_validator(pre=True)
def validate_pipeline_config(cls, values):
pipeline_config = values.get("pipeline_config", None)
if not pipeline_config:
raise ValueError("No 'pipeline_config' provided.")
if len(values) != 1:
raise ValueError(
"Only 'pipeline_config' key allowed when creating a pipeline structure object."
)
_config: PipelineConfig = pipeline_config
_steps: List[PipelineStep] = list(_config.steps)
_input_aliases: Dict[str, str] = dict(_config.input_aliases)
_output_aliases: Dict[str, str] = dict(_config.output_aliases)
values["steps"] = _steps
values["input_aliases"] = _input_aliases
values["output_aliases"] = _output_aliases
return values
generate_pipeline_endpoint_name(step_id, value_name)
¶Source code in kiara/models/module/pipeline/structure.py
def generate_pipeline_endpoint_name(step_id: str, value_name: str):
return f"{step_id}__{value_name}"
value_refs
¶
PipelineInputRef (ValueRef)
pydantic-model
¶An input to a pipeline.
Source code in kiara/models/module/pipeline/value_refs.py
class PipelineInputRef(ValueRef):
"""An input to a pipeline."""
connected_inputs: List[StepValueAddress] = Field(
description="The step inputs that are connected to this pipeline input",
default_factory=list,
)
is_constant: bool = Field(
"Whether this input is a constant and can't be changed by the user."
)
@property
def alias(self) -> str:
return generate_step_alias(PIPELINE_PARENT_MARKER, self.value_name)
PipelineOutputRef (ValueRef)
pydantic-model
¶An output to a pipeline.
Source code in kiara/models/module/pipeline/value_refs.py
class PipelineOutputRef(ValueRef):
"""An output to a pipeline."""
connected_output: StepValueAddress = Field(description="Connected step outputs.")
@property
def alias(self) -> str:
return generate_step_alias(PIPELINE_PARENT_MARKER, self.value_name)
StepInputRef (ValueRef)
pydantic-model
¶An input to a step.
This object can either have a 'connected_outputs' set, or a 'connected_pipeline_input', not both.
Source code in kiara/models/module/pipeline/value_refs.py
class StepInputRef(ValueRef):
"""An input to a step.
This object can either have a 'connected_outputs' set, or a 'connected_pipeline_input', not both.
"""
step_id: str = Field(description="The step id.")
connected_outputs: Optional[List[StepValueAddress]] = Field(
default=None,
description="A potential connected list of one or several module outputs.",
)
connected_pipeline_input: Optional[str] = Field(
default=None, description="A potential pipeline input."
)
is_constant: bool = Field(
"Whether this input is a constant and can't be changed by the user."
)
@root_validator(pre=True)
def ensure_single_connected_item(cls, values):
if values.get("connected_outputs", None) and values.get(
"connected_pipeline_input"
):
raise ValueError("Multiple connected items, only one allowed.")
return values
@property
def alias(self) -> str:
return generate_step_alias(self.step_id, self.value_name)
@property
def address(self) -> StepValueAddress:
return StepValueAddress(step_id=self.step_id, value_name=self.value_name)
def __str__(self):
name = camel_case_to_snake_case(self.__class__.__name__[0:-5], repl=" ")
return f"{name}: {self.step_id}.{self.value_name} ({self.value_schema.type})"
address: StepValueAddress
property
readonly
¶alias: str
property
readonly
¶connected_outputs: List[kiara.models.module.pipeline.value_refs.StepValueAddress]
pydantic-field
¶A potential connected list of one or several module outputs.
connected_pipeline_input: str
pydantic-field
¶A potential pipeline input.
is_constant: bool
pydantic-field
¶step_id: str
pydantic-field
required
¶The step id.
ensure_single_connected_item(values)
classmethod
¶Source code in kiara/models/module/pipeline/value_refs.py
@root_validator(pre=True)
def ensure_single_connected_item(cls, values):
if values.get("connected_outputs", None) and values.get(
"connected_pipeline_input"
):
raise ValueError("Multiple connected items, only one allowed.")
return values
StepOutputRef (ValueRef)
pydantic-model
¶An output to a step.
Source code in kiara/models/module/pipeline/value_refs.py
class StepOutputRef(ValueRef):
"""An output to a step."""
class Config:
allow_mutation = True
step_id: str = Field(description="The step id.")
pipeline_output: Optional[str] = Field(description="The connected pipeline output.")
connected_inputs: List[StepValueAddress] = Field(
description="The step inputs that are connected to this step output",
default_factory=list,
)
@property
def alias(self) -> str:
return generate_step_alias(self.step_id, self.value_name)
@property
def address(self) -> StepValueAddress:
return StepValueAddress(step_id=self.step_id, value_name=self.value_name)
def __str__(self):
name = camel_case_to_snake_case(self.__class__.__name__[0:-5], repl=" ")
return f"{name}: {self.step_id}.{self.value_name} ({self.value_schema.type})"
address: StepValueAddress
property
readonly
¶alias: str
property
readonly
¶connected_inputs: List[kiara.models.module.pipeline.value_refs.StepValueAddress]
pydantic-field
¶The step inputs that are connected to this step output
pipeline_output: str
pydantic-field
¶The connected pipeline output.
step_id: str
pydantic-field
required
¶The step id.
Config
¶Source code in kiara/models/module/pipeline/value_refs.py
class Config:
allow_mutation = True
StepValueAddress (BaseModel)
pydantic-model
¶Small model to describe the address of a value of a step, within a Pipeline/PipelineStructure.
Source code in kiara/models/module/pipeline/value_refs.py
class StepValueAddress(BaseModel):
"""Small model to describe the address of a value of a step, within a Pipeline/PipelineStructure."""
class Config:
extra = Extra.forbid
step_id: str = Field(description="The id of a step within a pipeline.")
value_name: str = Field(
description="The name of the value (output name or pipeline input name)."
)
sub_value: Optional[Dict[str, Any]] = Field(
default=None,
description="A reference to a subitem of a value (e.g. column, list item)",
)
@property
def alias(self):
"""An alias string for this address (in the form ``[step_id].[value_name]``)."""
return generate_step_alias(self.step_id, self.value_name)
def __eq__(self, other):
if not isinstance(other, StepValueAddress):
return False
return (self.step_id, self.value_name, self.sub_value) == (
other.step_id,
other.value_name,
other.sub_value,
)
def __hash__(self):
return hash((self.step_id, self.value_name, self.sub_value))
def __repr__(self):
if self.sub_value:
sub_value = f" sub_value={self.sub_value}"
else:
sub_value = ""
return f"{self.__class__.__name__}(step_id={self.step_id}, value_name={self.value_name}{sub_value})"
def __str__(self):
return self.__repr__()
alias
property
readonly
¶An alias string for this address (in the form [step_id].[value_name]).
step_id: str
pydantic-field
required
¶The id of a step within a pipeline.
sub_value: Dict[str, Any]
pydantic-field
¶A reference to a subitem of a value (e.g. column, list item)
value_name: str
pydantic-field
required
¶The name of the value (output name or pipeline input name).
Config
¶Source code in kiara/models/module/pipeline/value_refs.py
class Config:
extra = Extra.forbid
ValueRef (BaseModel)
pydantic-model
¶An object that holds information about the location of a value within a pipeline (or other structure).
Basically, a ValueRef helps the containing object where in its structure the value belongs (for example so
it can update dependent other values). A ValueRef object (obviously) does not contain the value itself.
There are four different ValueRef type that are relevant for pipelines:
- [kiara.pipeline.values.StepInputRef][]: an input to a step
- [kiara.pipeline.values.StepOutputRef][]: an output of a step
- [kiara.pipeline.values.PipelineInputRef][]: an input to a pipeline
- [kiara.pipeline.values.PipelineOutputRef][]: an output for a pipeline
Several ValueRef objects can target the same value, for example a step output and a connected step input would
reference the same Value (in most cases)..
Source code in kiara/models/module/pipeline/value_refs.py
class ValueRef(BaseModel):
"""An object that holds information about the location of a value within a pipeline (or other structure).
Basically, a `ValueRef` helps the containing object where in its structure the value belongs (for example so
it can update dependent other values). A `ValueRef` object (obviously) does not contain the value itself.
There are four different ValueRef type that are relevant for pipelines:
- [kiara.pipeline.values.StepInputRef][]: an input to a step
- [kiara.pipeline.values.StepOutputRef][]: an output of a step
- [kiara.pipeline.values.PipelineInputRef][]: an input to a pipeline
- [kiara.pipeline.values.PipelineOutputRef][]: an output for a pipeline
Several `ValueRef` objects can target the same value, for example a step output and a connected step input would
reference the same `Value` (in most cases)..
"""
class Config:
allow_mutation = True
extra = Extra.forbid
_id: uuid.UUID = PrivateAttr(default_factory=uuid.uuid4)
value_name: str
value_schema: ValueSchema
# pipeline_id: str
def __eq__(self, other):
if not isinstance(other, self.__class__):
return False
return self._id == other._id
def __hash__(self):
return hash(self._id)
def __repr__(self):
step_id = ""
if hasattr(self, "step_id"):
step_id = f" step_id='{self.step_id}'"
return f"{self.__class__.__name__}(value_name='{self.value_name}' {step_id})"
def __str__(self):
name = camel_case_to_snake_case(self.__class__.__name__[0:-5], repl=" ")
return f"{name}: {self.value_name} ({self.value_schema.type})"
generate_step_alias(step_id, value_name)
¶Source code in kiara/models/module/pipeline/value_refs.py
def generate_step_alias(step_id: str, value_name):
return f"{step_id}.{value_name}"
python_class
¶
Classes¶
PythonClass (KiaraModel)
pydantic-model
¶Python class and module information.
Source code in kiara/models/python_class.py
class PythonClass(KiaraModel):
"""Python class and module information."""
@classmethod
def from_class(cls, item_cls: Type, attach_context_metadata: bool = False):
cls_name = item_cls.__name__
module_name = item_cls.__module__
if module_name == "builtins":
full_name = cls_name
else:
full_name = f"{item_cls.__module__}.{item_cls.__name__}"
conf: Dict[str, Any] = {
"class_name": cls_name,
"module_name": module_name,
"full_name": full_name,
}
if attach_context_metadata:
ctx_md = ContextMetadataModel.from_class(item_cls=item_cls)
conf["items"] = ctx_md
result = PythonClass.construct(**conf)
result._cls_cache = item_cls
return result
class_name: str = Field(description="The name of the Python class.")
module_name: str = Field(
description="The name of the Python module this class lives in."
)
full_name: str = Field(description="The full class namespace.")
# context_metadata: Optional[ContextMetadataModel] = Field(
# description="Context metadata for the class.", default=None
# )
_module_cache: ModuleType = PrivateAttr(default=None)
_cls_cache: Type = PrivateAttr(default=None)
def _retrieve_id(self) -> str:
return self.full_name
def _retrieve_category_id(self) -> str:
return "metadata.python_class"
def _retrieve_data_to_hash(self) -> Any:
return self.full_name
def get_class(self) -> Type:
if self._cls_cache is None:
m = self.get_python_module()
self._cls_cache = getattr(m, self.class_name)
return self._cls_cache
def get_python_module(self) -> ModuleType:
if self._module_cache is None:
self._module_cache = importlib.import_module(self.module_name)
return self._module_cache
class_name: str
pydantic-field
required
¶The name of the Python class.
full_name: str
pydantic-field
required
¶The full class namespace.
module_name: str
pydantic-field
required
¶The name of the Python module this class lives in.
from_class(item_cls, attach_context_metadata=False)
classmethod
¶Source code in kiara/models/python_class.py
@classmethod
def from_class(cls, item_cls: Type, attach_context_metadata: bool = False):
cls_name = item_cls.__name__
module_name = item_cls.__module__
if module_name == "builtins":
full_name = cls_name
else:
full_name = f"{item_cls.__module__}.{item_cls.__name__}"
conf: Dict[str, Any] = {
"class_name": cls_name,
"module_name": module_name,
"full_name": full_name,
}
if attach_context_metadata:
ctx_md = ContextMetadataModel.from_class(item_cls=item_cls)
conf["items"] = ctx_md
result = PythonClass.construct(**conf)
result._cls_cache = item_cls
return result
get_class(self)
¶Source code in kiara/models/python_class.py
def get_class(self) -> Type:
if self._cls_cache is None:
m = self.get_python_module()
self._cls_cache = getattr(m, self.class_name)
return self._cls_cache
get_python_module(self)
¶Source code in kiara/models/python_class.py
def get_python_module(self) -> ModuleType:
if self._module_cache is None:
self._module_cache = importlib.import_module(self.module_name)
return self._module_cache
runtime_environment
special
¶
logger
¶
RuntimeEnvironment (KiaraModel)
pydantic-model
¶Source code in kiara/models/runtime_environment/__init__.py
class RuntimeEnvironment(KiaraModel):
class Config:
underscore_attrs_are_private = False
allow_mutation = False
@classmethod
def get_environment_type_name(cls) -> str:
env_type = cls.__fields__["environment_type"]
args = get_args(env_type.type_)
assert len(args) == 1
return args[0]
@classmethod
def create_environment_model(cls):
try:
type_name = cls.get_environment_type_name()
data = cls.retrieve_environment_data()
assert (
"environment_type" not in data.keys()
or data["environment_keys"] == type_name
)
data["environment_type"] = type_name
except Exception as e:
raise Exception(f"Can't create environment model for '{cls.__name__}': {e}")
return cls(**data)
def get_category_alias(self) -> str:
return f"{ENVIRONMENT_TYPE_CATEGORY_ID}.{self.environment_type}" # type: ignore
@classmethod
@abstractmethod
def retrieve_environment_data(cls) -> Dict[str, Any]:
pass
def _create_renderable_for_field(
self, field_name: str, for_summary: bool = False
) -> Optional[RenderableType]:
return extract_renderable(getattr(self, field_name))
def _retrieve_id(self) -> str:
return self.__class__.get_environment_type_name()
def _retrieve_category_id(self) -> str:
return ENVIRONMENT_TYPE_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
def create_renderable(self, **config: Any) -> RenderableType:
summary = config.get("summary", False)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("field")
table.add_column("summary")
for field_name, field in self.__fields__.items():
summary_item = self._create_renderable_for_field(
field_name, for_summary=summary
)
if summary_item is not None:
table.add_row(field_name, summary_item)
return table
Config
¶create_environment_model()
classmethod
¶Source code in kiara/models/runtime_environment/__init__.py
@classmethod
def create_environment_model(cls):
try:
type_name = cls.get_environment_type_name()
data = cls.retrieve_environment_data()
assert (
"environment_type" not in data.keys()
or data["environment_keys"] == type_name
)
data["environment_type"] = type_name
except Exception as e:
raise Exception(f"Can't create environment model for '{cls.__name__}': {e}")
return cls(**data)
create_renderable(self, **config)
¶Source code in kiara/models/runtime_environment/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
summary = config.get("summary", False)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("field")
table.add_column("summary")
for field_name, field in self.__fields__.items():
summary_item = self._create_renderable_for_field(
field_name, for_summary=summary
)
if summary_item is not None:
table.add_row(field_name, summary_item)
return table
get_category_alias(self)
¶Source code in kiara/models/runtime_environment/__init__.py
def get_category_alias(self) -> str:
return f"{ENVIRONMENT_TYPE_CATEGORY_ID}.{self.environment_type}" # type: ignore
get_environment_type_name()
classmethod
¶Source code in kiara/models/runtime_environment/__init__.py
@classmethod
def get_environment_type_name(cls) -> str:
env_type = cls.__fields__["environment_type"]
args = get_args(env_type.type_)
assert len(args) == 1
return args[0]
retrieve_environment_data()
classmethod
¶Source code in kiara/models/runtime_environment/__init__.py
@classmethod
@abstractmethod
def retrieve_environment_data(cls) -> Dict[str, Any]:
pass
Modules¶
kiara
¶
ArchiveTypeClassesInfo (TypeInfoModelGroup)
pydantic-model
¶Source code in kiara/models/runtime_environment/kiara.py
class ArchiveTypeClassesInfo(TypeInfoModelGroup):
@classmethod
def base_info_class(cls) -> Type[ArchiveTypeInfoModel]:
return ArchiveTypeInfoModel
type_name: Literal["archive_type"] = "archive_type"
type_infos: Mapping[str, ArchiveTypeInfoModel] = Field(
description="The archive info instances for each type."
)
type_infos: Mapping[str, kiara.models.runtime_environment.kiara.ArchiveTypeInfoModel]
pydantic-field
required
¶The archive info instances for each type.
type_name: Literal['archive_type']
pydantic-field
¶base_info_class()
classmethod
¶Source code in kiara/models/runtime_environment/kiara.py
@classmethod
def base_info_class(cls) -> Type[ArchiveTypeInfoModel]:
return ArchiveTypeInfoModel
ArchiveTypeInfoModel (KiaraTypeInfoModel)
pydantic-model
¶Source code in kiara/models/runtime_environment/kiara.py
class ArchiveTypeInfoModel(KiaraTypeInfoModel):
@classmethod
def create_from_type_class(
self, type_cls: Type[KiaraArchive]
) -> "ArchiveTypeInfoModel":
authors_md = AuthorsMetadataModel.from_class(type_cls)
doc = DocumentationMetadataModel.from_class_doc(type_cls)
python_class = PythonClass.from_class(type_cls)
properties_md = ContextMetadataModel.from_class(type_cls)
type_name = type_cls._archive_type_name # type: ignore
return ArchiveTypeInfoModel.construct(
type_name=type_name,
documentation=doc,
authors=authors_md,
context=properties_md,
python_class=python_class,
)
@classmethod
def base_class(self) -> Type[KiaraArchive]:
return KiaraArchive
@classmethod
def category_name(cls) -> str:
return "archive_type"
is_writable: bool = Field(
description="Whether this archive is writeable.", default=False
)
supported_item_types: List[str] = Field(
description="The item types this archive suports."
)
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
table.add_row("Python class", self.python_class.create_renderable())
table.add_row("is_writeable", "yes" if self.is_writable else "no")
table.add_row(
"supported_item_types", ", ".join(sorted(self.supported_item_types))
)
return table
is_writable: bool
pydantic-field
¶Whether this archive is writeable.
supported_item_types: List[str]
pydantic-field
required
¶The item types this archive suports.
base_class()
classmethod
¶Source code in kiara/models/runtime_environment/kiara.py
@classmethod
def base_class(self) -> Type[KiaraArchive]:
return KiaraArchive
category_name()
classmethod
¶Source code in kiara/models/runtime_environment/kiara.py
@classmethod
def category_name(cls) -> str:
return "archive_type"
create_from_type_class(type_cls)
classmethod
¶Source code in kiara/models/runtime_environment/kiara.py
@classmethod
def create_from_type_class(
self, type_cls: Type[KiaraArchive]
) -> "ArchiveTypeInfoModel":
authors_md = AuthorsMetadataModel.from_class(type_cls)
doc = DocumentationMetadataModel.from_class_doc(type_cls)
python_class = PythonClass.from_class(type_cls)
properties_md = ContextMetadataModel.from_class(type_cls)
type_name = type_cls._archive_type_name # type: ignore
return ArchiveTypeInfoModel.construct(
type_name=type_name,
documentation=doc,
authors=authors_md,
context=properties_md,
python_class=python_class,
)
create_renderable(self, **config)
¶Source code in kiara/models/runtime_environment/kiara.py
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
table.add_row("Python class", self.python_class.create_renderable())
table.add_row("is_writeable", "yes" if self.is_writable else "no")
table.add_row(
"supported_item_types", ", ".join(sorted(self.supported_item_types))
)
return table
KiaraTypesRuntimeEnvironment (RuntimeEnvironment)
pydantic-model
¶Source code in kiara/models/runtime_environment/kiara.py
class KiaraTypesRuntimeEnvironment(RuntimeEnvironment):
environment_type: Literal["kiara_types"]
archive_types: ArchiveTypeClassesInfo = Field(
description="The available implemented store types."
)
metadata_types: MetadataTypeClassesInfo = Field(
description="The available metadata types."
)
@classmethod
def retrieve_environment_data(cls) -> Dict[str, Any]:
result: Dict[str, Any] = {}
result["metadata_types"] = find_metadata_models()
result["archive_types"] = find_archive_types()
return result
archive_types: ArchiveTypeClassesInfo
pydantic-field
required
¶The available implemented store types.
environment_type: Literal['kiara_types']
pydantic-field
required
¶metadata_types: MetadataTypeClassesInfo
pydantic-field
required
¶The available metadata types.
retrieve_environment_data()
classmethod
¶Source code in kiara/models/runtime_environment/kiara.py
@classmethod
def retrieve_environment_data(cls) -> Dict[str, Any]:
result: Dict[str, Any] = {}
result["metadata_types"] = find_metadata_models()
result["archive_types"] = find_archive_types()
return result
find_archive_types(alias=None, only_for_package=None)
¶Source code in kiara/models/runtime_environment/kiara.py
def find_archive_types(
alias: Optional[str] = None, only_for_package: Optional[str] = None
) -> ArchiveTypeClassesInfo:
archive_types = find_all_archive_types()
group: ArchiveTypeClassesInfo = ArchiveTypeClassesInfo.create_from_type_items( # type: ignore
group_alias=alias, **archive_types
)
if only_for_package:
temp: Dict[str, KiaraTypeInfoModel] = {}
for key, info in group.items():
if info.context.labels.get("package") == only_for_package:
temp[key] = info # type: ignore
group = ArchiveTypeClassesInfo.construct(
group_id=group.group_id, group_alias=group.group_alias, type_infos=temp # type: ignore
)
return group
operating_system
¶
OSRuntimeEnvironment (RuntimeEnvironment)
pydantic-model
¶Manages information about the OS this kiara instance is running in.
TODO: details for other OS's (mainly BSDs)¶
Source code in kiara/models/runtime_environment/operating_system.py
class OSRuntimeEnvironment(RuntimeEnvironment):
"""Manages information about the OS this kiara instance is running in.
# TODO: details for other OS's (mainly BSDs)
"""
environment_type: typing.Literal["operating_system"]
operation_system: str = Field(description="The operation system name.")
platform: str = Field(description="The platform name.")
release: str = Field(description="The platform release name.")
version: str = Field(description="The platform version name.")
machine: str = Field(description="The architecture.")
os_specific: typing.Dict[str, typing.Any] = Field(
description="OS specific platform metadata.", default_factory=dict
)
@classmethod
def retrieve_environment_data(self) -> typing.Dict[str, typing.Any]:
os_specific: typing.Dict[str, typing.Any] = {}
platform_system = platform.system()
if platform_system == "Linux":
import distro
os_specific["distribution"] = {
"name": distro.name(),
"version": distro.version(),
"codename": distro.codename(),
}
elif platform_system == "Darwin":
mac_version = platform.mac_ver()
os_specific["mac_ver_release"] = mac_version[0]
os_specific["mac_ver_machine"] = mac_version[2]
result = {
"operation_system": os.name,
"platform": platform_system,
"release": platform.release(),
"version": platform.version(),
"machine": platform.machine(),
"os_specific": os_specific,
}
# if config.include_all_info:
# result["uname"] = platform.uname()._asdict()
return result
environment_type: Literal['operating_system']
pydantic-field
required
¶machine: str
pydantic-field
required
¶The architecture.
operation_system: str
pydantic-field
required
¶The operation system name.
os_specific: Dict[str, Any]
pydantic-field
¶OS specific platform metadata.
platform: str
pydantic-field
required
¶The platform name.
release: str
pydantic-field
required
¶The platform release name.
version: str
pydantic-field
required
¶The platform version name.
retrieve_environment_data()
classmethod
¶Source code in kiara/models/runtime_environment/operating_system.py
@classmethod
def retrieve_environment_data(self) -> typing.Dict[str, typing.Any]:
os_specific: typing.Dict[str, typing.Any] = {}
platform_system = platform.system()
if platform_system == "Linux":
import distro
os_specific["distribution"] = {
"name": distro.name(),
"version": distro.version(),
"codename": distro.codename(),
}
elif platform_system == "Darwin":
mac_version = platform.mac_ver()
os_specific["mac_ver_release"] = mac_version[0]
os_specific["mac_ver_machine"] = mac_version[2]
result = {
"operation_system": os.name,
"platform": platform_system,
"release": platform.release(),
"version": platform.version(),
"machine": platform.machine(),
"os_specific": os_specific,
}
# if config.include_all_info:
# result["uname"] = platform.uname()._asdict()
return result
python
¶
PythonPackage (BaseModel)
pydantic-model
¶Source code in kiara/models/runtime_environment/python.py
class PythonPackage(BaseModel):
name: str = Field(description="The name of the Python package.")
version: str = Field(description="The version of the package.")
PythonRuntimeEnvironment (RuntimeEnvironment)
pydantic-model
¶Source code in kiara/models/runtime_environment/python.py
class PythonRuntimeEnvironment(RuntimeEnvironment):
environment_type: Literal["python"]
python_version: str = Field(description="The version of Python.")
packages: List[PythonPackage] = Field(
description="The packages installed in the Python (virtual) environment."
)
# python_config: typing.Dict[str, str] = Field(
# description="Configuration details about the Python installation."
# )
def _create_renderable_for_field(
self, field_name: str, for_summary: bool = False
) -> Optional[RenderableType]:
if field_name != "packages":
return extract_renderable(getattr(self, field_name))
if for_summary:
return ", ".join(p.name for p in self.packages)
table = Table(show_header=True, box=box.SIMPLE)
table.add_column("package name")
table.add_column("version")
for package in self.packages:
table.add_row(package.name, package.version)
return table
@classmethod
def retrieve_environment_data(cls) -> Dict[str, Any]:
packages = []
all_packages = packages_distributions()
for name, pkgs in all_packages.items():
for pkg in pkgs:
dist = distribution(pkg)
packages.append({"name": name, "version": dist.version})
result: Dict[str, Any] = {
"python_version": sys.version,
"packages": sorted(packages, key=lambda x: x["name"]),
}
# if config.include_all_info:
# import sysconfig
# result["python_config"] = sysconfig.get_config_vars()
return result
environment_type: Literal['python']
pydantic-field
required
¶packages: List[kiara.models.runtime_environment.python.PythonPackage]
pydantic-field
required
¶The packages installed in the Python (virtual) environment.
python_version: str
pydantic-field
required
¶The version of Python.
retrieve_environment_data()
classmethod
¶Source code in kiara/models/runtime_environment/python.py
@classmethod
def retrieve_environment_data(cls) -> Dict[str, Any]:
packages = []
all_packages = packages_distributions()
for name, pkgs in all_packages.items():
for pkg in pkgs:
dist = distribution(pkg)
packages.append({"name": name, "version": dist.version})
result: Dict[str, Any] = {
"python_version": sys.version,
"packages": sorted(packages, key=lambda x: x["name"]),
}
# if config.include_all_info:
# import sysconfig
# result["python_config"] = sysconfig.get_config_vars()
return result
values
special
¶
Classes¶
ValueStatus (Enum)
¶Modules¶
data_type
¶
DataTypeClassInfo (KiaraTypeInfoModel)
pydantic-model
¶Source code in kiara/models/values/data_type.py
class DataTypeClassInfo(KiaraTypeInfoModel[DataType]):
@classmethod
def create_from_type_class(
self, type_cls: Type[DataType], kiara: Optional["Kiara"] = None
) -> "DataTypeClassInfo":
authors = AuthorsMetadataModel.from_class(type_cls)
doc = DocumentationMetadataModel.from_class_doc(type_cls)
properties_md = ContextMetadataModel.from_class(type_cls)
if kiara is not None:
qual_profiles = kiara.type_registry.get_associated_profiles(type_cls._data_type_name) # type: ignore
lineage = kiara.type_registry.get_type_lineage(type_cls._data_type_name) # type: ignore
else:
qual_profiles = None
lineage = None
try:
result = DataTypeClassInfo.construct(
type_name=type_cls._data_type_name, # type: ignore
python_class=PythonClass.from_class(type_cls),
value_cls=PythonClass.from_class(type_cls.python_class()),
data_type_config_cls=PythonClass.from_class(
type_cls.data_type_config_class()
),
lineage=lineage, # type: ignore
qualifier_profiles=qual_profiles,
documentation=doc,
authors=authors,
context=properties_md,
)
except Exception as e:
if isinstance(
e, TypeError
) and "missing 1 required positional argument: 'cls'" in str(e):
raise Exception(
f"Invalid implementation of TypeValue subclass '{type_cls.__name__}': 'python_class' method must be marked as a '@classmethod'. This is a bug."
)
raise e
result._kiara = kiara
return result
@classmethod
def base_class(self) -> Type[DataType]:
return DataType
@classmethod
def category_name(cls) -> str:
return "data_type"
value_cls: PythonClass = Field(description="The python class of the value itself.")
data_type_config_cls: PythonClass = Field(
description="The python class holding the schema for configuring this type."
)
lineage: Optional[List[str]] = Field(description="This types lineage.")
qualifier_profiles: Optional[Mapping[str, Mapping[str, Any]]] = Field(
description="A map of qualifier profiles for this data types."
)
_kiara: Optional["Kiara"] = PrivateAttr(default=None)
def _retrieve_id(self) -> str:
return self.type_name
def _retrieve_category_id(self) -> str:
return DATA_TYPE_CLASS_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.type_name
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if self.lineage:
table.add_row("lineage", "\n".join(self.lineage[0:]))
else:
table.add_row("lineage", "-- n/a --")
if self.qualifier_profiles:
qual_table = Table(show_header=False, box=box.SIMPLE)
qual_table.add_column("name")
qual_table.add_column("config")
for name, details in self.qualifier_profiles.items():
json_details = orjson_dumps(details, option=orjson.OPT_INDENT_2)
qual_table.add_row(
name, Syntax(json_details, "json", background_color="default")
)
table.add_row("qualifier profile(s)", qual_table)
else:
table.add_row("qualifier profile(s)", "-- n/a --")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
table.add_row("Python class", self.python_class.create_renderable())
table.add_row("Config class", self.data_type_config_cls.create_renderable())
table.add_row("Value class", self.value_cls.create_renderable())
return table
data_type_config_cls: PythonClass
pydantic-field
required
¶The python class holding the schema for configuring this type.
lineage: List[str]
pydantic-field
¶This types lineage.
qualifier_profiles: Mapping[str, Mapping[str, Any]]
pydantic-field
¶A map of qualifier profiles for this data types.
value_cls: PythonClass
pydantic-field
required
¶The python class of the value itself.
base_class()
classmethod
¶Source code in kiara/models/values/data_type.py
@classmethod
def base_class(self) -> Type[DataType]:
return DataType
category_name()
classmethod
¶Source code in kiara/models/values/data_type.py
@classmethod
def category_name(cls) -> str:
return "data_type"
create_from_type_class(type_cls, kiara=None)
classmethod
¶Source code in kiara/models/values/data_type.py
@classmethod
def create_from_type_class(
self, type_cls: Type[DataType], kiara: Optional["Kiara"] = None
) -> "DataTypeClassInfo":
authors = AuthorsMetadataModel.from_class(type_cls)
doc = DocumentationMetadataModel.from_class_doc(type_cls)
properties_md = ContextMetadataModel.from_class(type_cls)
if kiara is not None:
qual_profiles = kiara.type_registry.get_associated_profiles(type_cls._data_type_name) # type: ignore
lineage = kiara.type_registry.get_type_lineage(type_cls._data_type_name) # type: ignore
else:
qual_profiles = None
lineage = None
try:
result = DataTypeClassInfo.construct(
type_name=type_cls._data_type_name, # type: ignore
python_class=PythonClass.from_class(type_cls),
value_cls=PythonClass.from_class(type_cls.python_class()),
data_type_config_cls=PythonClass.from_class(
type_cls.data_type_config_class()
),
lineage=lineage, # type: ignore
qualifier_profiles=qual_profiles,
documentation=doc,
authors=authors,
context=properties_md,
)
except Exception as e:
if isinstance(
e, TypeError
) and "missing 1 required positional argument: 'cls'" in str(e):
raise Exception(
f"Invalid implementation of TypeValue subclass '{type_cls.__name__}': 'python_class' method must be marked as a '@classmethod'. This is a bug."
)
raise e
result._kiara = kiara
return result
create_renderable(self, **config)
¶Source code in kiara/models/values/data_type.py
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if self.lineage:
table.add_row("lineage", "\n".join(self.lineage[0:]))
else:
table.add_row("lineage", "-- n/a --")
if self.qualifier_profiles:
qual_table = Table(show_header=False, box=box.SIMPLE)
qual_table.add_column("name")
qual_table.add_column("config")
for name, details in self.qualifier_profiles.items():
json_details = orjson_dumps(details, option=orjson.OPT_INDENT_2)
qual_table.add_row(
name, Syntax(json_details, "json", background_color="default")
)
table.add_row("qualifier profile(s)", qual_table)
else:
table.add_row("qualifier profile(s)", "-- n/a --")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
table.add_row("Python class", self.python_class.create_renderable())
table.add_row("Config class", self.data_type_config_cls.create_renderable())
table.add_row("Value class", self.value_cls.create_renderable())
return table
DataTypeClassesInfo (TypeInfoModelGroup)
pydantic-model
¶Source code in kiara/models/values/data_type.py
class DataTypeClassesInfo(TypeInfoModelGroup):
@classmethod
def create_from_type_items(
cls,
group_alias: Optional[str] = None,
**items: Type,
) -> "TypeInfoModelGroup":
type_infos = {
k: cls.base_info_class().create_from_type_class(v) for k, v in items.items() # type: ignore
}
data_types_info = cls.construct(group_alias=group_alias, type_infos=type_infos) # type: ignore
return data_types_info
@classmethod
def create_augmented_from_type_items(
cls,
kiara: Optional["Kiara"] = None,
group_alias: Optional[str] = None,
**items: Type,
) -> "TypeInfoModelGroup":
type_infos = {
k: cls.base_info_class().create_from_type_class(v, kiara=kiara) for k, v in items.items() # type: ignore
}
data_types_info = cls.construct(group_alias=group_alias, type_infos=type_infos) # type: ignore
data_types_info._kiara = kiara
return data_types_info
@classmethod
def base_info_class(cls) -> Type[DataTypeClassInfo]:
return DataTypeClassInfo
type_name: Literal["data_type"] = "data_type"
type_infos: Mapping[str, DataTypeClassInfo] = Field(
description="The data_type info instances for each type."
)
_kiara: Optional["Kiara"] = PrivateAttr(default=None)
def create_renderable(self, **config: Any) -> RenderableType:
full_doc = config.get("full_doc", False)
show_subtypes_inline = config.get("show_qualifier_profiles_inline", True)
show_lineage = config.get("show_type_lineage", True)
show_lines = full_doc or show_subtypes_inline or show_lineage
table = Table(show_header=True, box=box.SIMPLE, show_lines=show_lines)
table.add_column("type name", style="i")
if show_lineage:
table.add_column("type lineage")
if show_subtypes_inline:
table.add_column("(qualifier) profiles")
if full_doc:
table.add_column("documentation")
else:
table.add_column("description")
all_types = self.type_infos.keys()
for type_name in sorted(all_types): # type: ignore
t_md = self.type_infos[type_name] # type: ignore
row: List[Any] = [type_name]
if show_lineage:
if self._kiara is None:
lineage_str = "-- n/a --"
else:
lineage = list(
self._kiara.type_registry.get_type_lineage(type_name)
)
lineage_str = ", ".join(reversed(lineage[1:]))
row.append(lineage_str)
if show_subtypes_inline:
if self._kiara is None:
qual_profiles = "-- n/a --"
else:
qual_p = self._kiara.type_registry.get_associated_profiles(
data_type_name=type_name
).keys()
if qual_p:
qual_profiles = "\n".join(qual_p)
else:
qual_profiles = "-- n/a --"
row.append(qual_profiles)
if full_doc:
md = Markdown(t_md.documentation.full_doc)
else:
md = Markdown(t_md.documentation.description)
row.append(md)
table.add_row(*row)
return table
type_infos: Mapping[str, kiara.models.values.data_type.DataTypeClassInfo]
pydantic-field
required
¶The data_type info instances for each type.
type_name: Literal['data_type']
pydantic-field
¶base_info_class()
classmethod
¶Source code in kiara/models/values/data_type.py
@classmethod
def base_info_class(cls) -> Type[DataTypeClassInfo]:
return DataTypeClassInfo
create_augmented_from_type_items(kiara=None, group_alias=None, **items)
classmethod
¶Source code in kiara/models/values/data_type.py
@classmethod
def create_augmented_from_type_items(
cls,
kiara: Optional["Kiara"] = None,
group_alias: Optional[str] = None,
**items: Type,
) -> "TypeInfoModelGroup":
type_infos = {
k: cls.base_info_class().create_from_type_class(v, kiara=kiara) for k, v in items.items() # type: ignore
}
data_types_info = cls.construct(group_alias=group_alias, type_infos=type_infos) # type: ignore
data_types_info._kiara = kiara
return data_types_info
create_from_type_items(group_alias=None, **items)
classmethod
¶Source code in kiara/models/values/data_type.py
@classmethod
def create_from_type_items(
cls,
group_alias: Optional[str] = None,
**items: Type,
) -> "TypeInfoModelGroup":
type_infos = {
k: cls.base_info_class().create_from_type_class(v) for k, v in items.items() # type: ignore
}
data_types_info = cls.construct(group_alias=group_alias, type_infos=type_infos) # type: ignore
return data_types_info
create_renderable(self, **config)
¶Source code in kiara/models/values/data_type.py
def create_renderable(self, **config: Any) -> RenderableType:
full_doc = config.get("full_doc", False)
show_subtypes_inline = config.get("show_qualifier_profiles_inline", True)
show_lineage = config.get("show_type_lineage", True)
show_lines = full_doc or show_subtypes_inline or show_lineage
table = Table(show_header=True, box=box.SIMPLE, show_lines=show_lines)
table.add_column("type name", style="i")
if show_lineage:
table.add_column("type lineage")
if show_subtypes_inline:
table.add_column("(qualifier) profiles")
if full_doc:
table.add_column("documentation")
else:
table.add_column("description")
all_types = self.type_infos.keys()
for type_name in sorted(all_types): # type: ignore
t_md = self.type_infos[type_name] # type: ignore
row: List[Any] = [type_name]
if show_lineage:
if self._kiara is None:
lineage_str = "-- n/a --"
else:
lineage = list(
self._kiara.type_registry.get_type_lineage(type_name)
)
lineage_str = ", ".join(reversed(lineage[1:]))
row.append(lineage_str)
if show_subtypes_inline:
if self._kiara is None:
qual_profiles = "-- n/a --"
else:
qual_p = self._kiara.type_registry.get_associated_profiles(
data_type_name=type_name
).keys()
if qual_p:
qual_profiles = "\n".join(qual_p)
else:
qual_profiles = "-- n/a --"
row.append(qual_profiles)
if full_doc:
md = Markdown(t_md.documentation.full_doc)
else:
md = Markdown(t_md.documentation.description)
row.append(md)
table.add_row(*row)
return table
info
¶RENDER_FIELDS: Dict[str, Dict[str, Any]]
¶
ValueInfo (Value)
pydantic-model
¶Source code in kiara/models/values/info.py
class ValueInfo(Value):
@classmethod
def create_from_value(
cls,
kiara: "Kiara",
value: Value,
resolve_aliases: bool = True,
resolve_destinies: bool = True,
):
if resolve_aliases:
aliases = sorted(
kiara.alias_registry.find_aliases_for_value_id(value.value_id)
)
else:
aliases = None
if value.is_stored:
load_config = kiara.data_registry.retrieve_load_config(
value_id=value.value_id
)
else:
load_config = None
is_internal = "internal" in kiara.type_registry.get_type_lineage(
value.data_type_name
)
if resolve_destinies:
destiny_links = kiara.data_registry.find_destinies_for_value(
value_id=value.value_id
)
filtered_destinies = {}
for alias, value_id in destiny_links.items():
if (
alias in value.property_links.keys()
and value_id == value.property_links[alias]
):
continue
filtered_destinies[alias] = value_id
else:
filtered_destinies = None
model = ValueInfo.construct(
value_id=value.value_id,
kiara_id=value.kiara_id,
value_schema=value.value_schema,
value_status=value.value_status,
value_size=value.value_size,
value_hash=value.value_hash,
pedigree=value.pedigree,
pedigree_output_name=value.pedigree_output_name,
data_type_class=value.data_type_class,
property_links=value.property_links,
destiny_links=filtered_destinies,
destiny_backlinks=value.destiny_backlinks,
aliases=aliases,
load_config=load_config,
)
model._set_registry(value._data_registry)
model._alias_registry = kiara.alias_registry # type: ignore
model._is_stored = value._is_stored
model._data_type = value._data_type
model._value_data = value._value_data
model._data_retrieved = value._data_retrieved
model._is_internal = is_internal
return model
value_id: uuid.UUID = Field(description="The value id.")
value_schema: ValueSchema = Field(description="The data schema of this value.")
aliases: Optional[List[str]] = Field(
description="The aliases that are registered for this value."
)
load_config: Optional[LoadConfig] = Field(
description="The load config associated with this value."
)
destiny_links: Optional[Mapping[str, uuid.UUID]] = Field(
description="References to all the values that act as destiny for this value in this context."
)
_is_internal: bool = PrivateAttr(default=False)
_alias_registry: AliasRegistry = PrivateAttr(default=None)
def _retrieve_id(self) -> str:
return str(self.value_id)
def _retrieve_category_id(self) -> str:
return "instance.value_info"
def _retrieve_data_to_hash(self) -> Any:
return self.value_id
def resolve_aliases(self):
aliases = self._alias_registry.find_aliases_for_value_id(self.value_id)
if aliases:
aliases = sorted(aliases)
self.aliases = aliases
def resolve_destinies(self):
destiny_links = self._data_registry.find_destinies_for_value(
value_id=self.value_id
)
filtered_destinies = {}
for alias, value_id in destiny_links.items():
if (
alias in self.property_links.keys()
and value_id == self.property_links[alias]
):
continue
filtered_destinies[alias] = value_id
self.destiny_links = filtered_destinies
aliases: List[str]
pydantic-field
¶The aliases that are registered for this value.
destiny_links: Mapping[str, uuid.UUID]
pydantic-field
¶References to all the values that act as destiny for this value in this context.
load_config: LoadConfig
pydantic-field
¶The load config associated with this value.
create_from_value(kiara, value, resolve_aliases=True, resolve_destinies=True)
classmethod
¶Source code in kiara/models/values/info.py
@classmethod
def create_from_value(
cls,
kiara: "Kiara",
value: Value,
resolve_aliases: bool = True,
resolve_destinies: bool = True,
):
if resolve_aliases:
aliases = sorted(
kiara.alias_registry.find_aliases_for_value_id(value.value_id)
)
else:
aliases = None
if value.is_stored:
load_config = kiara.data_registry.retrieve_load_config(
value_id=value.value_id
)
else:
load_config = None
is_internal = "internal" in kiara.type_registry.get_type_lineage(
value.data_type_name
)
if resolve_destinies:
destiny_links = kiara.data_registry.find_destinies_for_value(
value_id=value.value_id
)
filtered_destinies = {}
for alias, value_id in destiny_links.items():
if (
alias in value.property_links.keys()
and value_id == value.property_links[alias]
):
continue
filtered_destinies[alias] = value_id
else:
filtered_destinies = None
model = ValueInfo.construct(
value_id=value.value_id,
kiara_id=value.kiara_id,
value_schema=value.value_schema,
value_status=value.value_status,
value_size=value.value_size,
value_hash=value.value_hash,
pedigree=value.pedigree,
pedigree_output_name=value.pedigree_output_name,
data_type_class=value.data_type_class,
property_links=value.property_links,
destiny_links=filtered_destinies,
destiny_backlinks=value.destiny_backlinks,
aliases=aliases,
load_config=load_config,
)
model._set_registry(value._data_registry)
model._alias_registry = kiara.alias_registry # type: ignore
model._is_stored = value._is_stored
model._data_type = value._data_type
model._value_data = value._value_data
model._data_retrieved = value._data_retrieved
model._is_internal = is_internal
return model
resolve_aliases(self)
¶Source code in kiara/models/values/info.py
def resolve_aliases(self):
aliases = self._alias_registry.find_aliases_for_value_id(self.value_id)
if aliases:
aliases = sorted(aliases)
self.aliases = aliases
resolve_destinies(self)
¶Source code in kiara/models/values/info.py
def resolve_destinies(self):
destiny_links = self._data_registry.find_destinies_for_value(
value_id=self.value_id
)
filtered_destinies = {}
for alias, value_id in destiny_links.items():
if (
alias in self.property_links.keys()
and value_id == self.property_links[alias]
):
continue
filtered_destinies[alias] = value_id
self.destiny_links = filtered_destinies
ValuesInfo (BaseModel)
pydantic-model
¶Source code in kiara/models/values/info.py
class ValuesInfo(BaseModel):
@classmethod
def create_from_values(cls, kiara: "Kiara", *values: Union[Value, uuid.UUID]):
v = [
ValueInfo.create_from_value(
kiara=kiara,
value=v if isinstance(v, Value) else kiara.data_registry.get_value(v),
)
for v in values
]
return ValuesInfo(__root__=v)
__root__: List[ValueInfo]
def create_render_map(self, render_type: str = "terminal", **render_config):
list_by_alias = render_config.get("list_by_alias", True)
show_internal = render_config.get("show_internal", False)
render_fields = render_config.get("render_fields", None)
if not render_fields:
render_fields = [k for k, v in RENDER_FIELDS.items() if v["show_default"]]
if list_by_alias:
render_fields[0] = "aliases"
render_fields[1] = "value_id"
render_map: Dict[uuid.UUID, Dict[str, Any]] = {}
lookup = {}
for value in self.__root__:
if not show_internal and value._is_internal:
continue
lookup[value.value_id] = value
details = {}
for property in render_fields:
if hasattr(value, property) and property != "data":
attr = getattr(value, property)
else:
attr = value
render_func = (
RENDER_FIELDS.get(property, {})
.get("render", {})
.get(render_type, None)
)
if render_func is None:
rendered = attr
else:
rendered = render_func(attr)
details[property] = rendered
render_map[value.value_id] = details
if not list_by_alias:
return {str(k): v for k, v in render_map.items()}
else:
result: Dict[str, Dict[str, Any]] = {}
for value_id, render_details in render_map.items():
value_aliases = lookup[value_id].aliases
if value_aliases:
for alias in value_aliases:
assert alias not in result.keys()
render_details = dict(render_details)
render_details["alias"] = alias
result[alias] = render_details
else:
render_details["alias"] = ""
result[f"no_aliases_{value_id}"] = render_details
return result
def create_renderable(self, render_type: str = "terminal", **render_config: Any):
render_map = self.create_render_map(render_type=render_type, **render_config)
list_by_alias = render_config.get("list_by_alias", True)
render_fields = render_config.get("render_fields", None)
if not render_fields:
render_fields = [k for k, v in RENDER_FIELDS.items() if v["show_default"]]
if list_by_alias:
render_fields.insert(0, "alias")
render_fields.remove("aliases")
table = Table(box=box.SIMPLE)
for property in render_fields:
if property == "aliases" and list_by_alias:
table.add_column("alias")
else:
table.add_column(property)
for item_id, details in render_map.items():
row = []
for field in render_fields:
value = details[field]
row.append(value)
table.add_row(*row)
return table
create_from_values(kiara, *values)
classmethod
¶Source code in kiara/models/values/info.py
@classmethod
def create_from_values(cls, kiara: "Kiara", *values: Union[Value, uuid.UUID]):
v = [
ValueInfo.create_from_value(
kiara=kiara,
value=v if isinstance(v, Value) else kiara.data_registry.get_value(v),
)
for v in values
]
return ValuesInfo(__root__=v)
create_render_map(self, render_type='terminal', **render_config)
¶Source code in kiara/models/values/info.py
def create_render_map(self, render_type: str = "terminal", **render_config):
list_by_alias = render_config.get("list_by_alias", True)
show_internal = render_config.get("show_internal", False)
render_fields = render_config.get("render_fields", None)
if not render_fields:
render_fields = [k for k, v in RENDER_FIELDS.items() if v["show_default"]]
if list_by_alias:
render_fields[0] = "aliases"
render_fields[1] = "value_id"
render_map: Dict[uuid.UUID, Dict[str, Any]] = {}
lookup = {}
for value in self.__root__:
if not show_internal and value._is_internal:
continue
lookup[value.value_id] = value
details = {}
for property in render_fields:
if hasattr(value, property) and property != "data":
attr = getattr(value, property)
else:
attr = value
render_func = (
RENDER_FIELDS.get(property, {})
.get("render", {})
.get(render_type, None)
)
if render_func is None:
rendered = attr
else:
rendered = render_func(attr)
details[property] = rendered
render_map[value.value_id] = details
if not list_by_alias:
return {str(k): v for k, v in render_map.items()}
else:
result: Dict[str, Dict[str, Any]] = {}
for value_id, render_details in render_map.items():
value_aliases = lookup[value_id].aliases
if value_aliases:
for alias in value_aliases:
assert alias not in result.keys()
render_details = dict(render_details)
render_details["alias"] = alias
result[alias] = render_details
else:
render_details["alias"] = ""
result[f"no_aliases_{value_id}"] = render_details
return result
create_renderable(self, render_type='terminal', **render_config)
¶Source code in kiara/models/values/info.py
def create_renderable(self, render_type: str = "terminal", **render_config: Any):
render_map = self.create_render_map(render_type=render_type, **render_config)
list_by_alias = render_config.get("list_by_alias", True)
render_fields = render_config.get("render_fields", None)
if not render_fields:
render_fields = [k for k, v in RENDER_FIELDS.items() if v["show_default"]]
if list_by_alias:
render_fields.insert(0, "alias")
render_fields.remove("aliases")
table = Table(box=box.SIMPLE)
for property in render_fields:
if property == "aliases" and list_by_alias:
table.add_column("alias")
else:
table.add_column(property)
for item_id, details in render_map.items():
row = []
for field in render_fields:
value = details[field]
row.append(value)
table.add_row(*row)
return table
render_value_data(value)
¶Source code in kiara/models/values/info.py
def render_value_data(value: Value):
try:
renderable = value._data_registry.render_data(
value.value_id, target_type="terminal_renderable"
)
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
log_message("error.render_value", value=value.value_id, error=e)
renderable = [str(value.data)]
return renderable
value
¶ORPHAN
¶log
¶yaml
¶
UnloadableData (KiaraModel)
pydantic-model
¶A special 'marker' model, indicating that the data of value can't be loaded.
In most cases, the reason this happens is because the current kiara context is missing some value types and/or modules.
Source code in kiara/models/values/value.py
class UnloadableData(KiaraModel):
"""A special 'marker' model, indicating that the data of value can't be loaded.
In most cases, the reason this happens is because the current kiara context is missing some value types and/or modules."""
value: Value = Field(description="A reference to the value.")
load_config: LoadConfig = Field(description="The load config")
def _retrieve_id(self) -> str:
return self.value.model_id
def _retrieve_category_id(self) -> str:
return UNOLOADABLE_DATA_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.value.model_data_hash
Value (ValueDetails)
pydantic-model
¶Source code in kiara/models/values/value.py
class Value(ValueDetails):
_value_data: Any = PrivateAttr(default=SpecialValue.NOT_SET)
_data_retrieved: bool = PrivateAttr(default=False)
_data_registry: "DataRegistry" = PrivateAttr(default=None)
_data_type: "DataType" = PrivateAttr(default=None)
_is_stored: bool = PrivateAttr(default=False)
_cached_properties: Optional["ValueMap"] = PrivateAttr(default=None)
property_links: Mapping[str, uuid.UUID] = Field(
description="Links to values that are properties of this value.",
default_factory=dict,
)
destiny_backlinks: Mapping[uuid.UUID, str] = Field(
description="Backlinks to values that this value acts as destiny/or property for.",
default_factory=dict,
)
def add_property(
self,
value_id: Union[uuid.UUID, "Value"],
property_path: str,
add_origin_to_property_value: bool = True,
):
value = None
try:
value_temp = value
value_id = value_id.value_id # type: ignore
value = value_temp
except Exception:
# in case a Value object was provided
pass
finally:
del value_temp
if add_origin_to_property_value:
if value is None:
value = self._data_registry.get_value(value_id=value_id) # type: ignore
if value._is_stored:
raise Exception(
f"Can't add property to value '{self.value_id}': referenced value '{value.value_id}' already locked, so it's not possible to add the property backlink (as requested)."
)
assert value is not None
if self._is_stored:
raise Exception(
f"Can't add property to value '{self.value_id}': value already locked."
)
if property_path in self.property_links.keys():
raise Exception(
f"Can't add property to value '{self.value_id}': property '{property_path}' already set."
)
self.property_links[property_path] = value_id # type: ignore
if add_origin_to_property_value:
value.add_destiny_details(
value_id=self.value_id, destiny_alias=property_path
)
self._cached_properties = None
def add_destiny_details(self, value_id: uuid.UUID, destiny_alias: str):
if self._is_stored:
raise Exception(
f"Can't set destiny_refs to value '{self.value_id}': value already locked."
)
self.destiny_backlinks[value_id] = destiny_alias # type: ignore
@property
def data(self) -> Any:
if not self.is_initialized:
raise Exception(
f"Can't retrieve data for value '{self.value_id}': value not initialized yet. This is most likely a bug."
)
return self._retrieve_data()
def _retrieve_data(self) -> Any:
if self._value_data is not SpecialValue.NOT_SET:
return self._value_data
if self.value_status in [ValueStatus.NOT_SET, ValueStatus.NONE]:
self._value_data = None
return self._value_data
elif self.value_status not in [ValueStatus.SET, ValueStatus.DEFAULT]:
raise Exception(f"Invalid internal state of value '{self.value_id}'.")
retrieved = self._data_registry.retrieve_value_data(value_id=self.value_id)
if retrieved is None or isinstance(retrieved, SpecialValue):
raise Exception(
f"Can't set value data, invalid data type: {type(retrieved)}"
)
self._value_data = retrieved
self._data_retrieved = True
return self._value_data
def retrieve_load_config(self) -> Optional[LoadConfig]:
return self._data_registry.retrieve_load_config(value_id=self.value_id)
def __repr__(self):
return f"{self.__class__.__name__}(id={self.value_id}, type={self.data_type_name}, status={self.value_status.value}, initialized={self.is_initialized} optional={self.value_schema.optional})"
def _set_registry(self, data_registry: "DataRegistry") -> None:
self._data_registry = data_registry
@property
def is_initialized(self) -> bool:
result = not self.is_set or self._data_registry is not None
return result
@property
def is_stored(self) -> bool:
return self._is_stored
@property
def data_type(self) -> "DataType":
if self._data_type is not None:
return self._data_type
self._data_type = self.data_type_class.get_class()(
**self.value_schema.type_config
)
return self._data_type
@property
def property_values(self) -> "ValueMap":
if self._cached_properties is not None:
return self._cached_properties
self._cached_properties = self._data_registry.load_values(self.property_links)
return self._cached_properties
@property
def property_names(self) -> Iterable[str]:
return self.property_links.keys()
def get_property_value(self, property_key) -> "Value":
if property_key not in self.property_links.keys():
raise Exception(
f"Value '{self.value_id}' has no property with key '{property_key}."
)
return self._data_registry.get_value(self.property_links[property_key])
def get_property_data(self, property_key: str) -> Any:
return self.get_property_value(property_key=property_key).data
def create_renderable(self, **render_config: Any) -> RenderableType:
from kiara.utils.output import extract_renderable
show_pedigree = render_config.get("show_pedigree", False)
show_load_config = render_config.get("show_load_config", False)
show_properties = render_config.get("show_properties", False)
show_destinies = render_config.get("show_destinies", False)
show_destiny_backlinks = render_config.get("show_destiny_backlinks", False)
show_data = render_config.get("show_data_preview", False)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("Key", style="i")
table.add_column("Value")
table.add_row("value_id", str(self.value_id))
if hasattr(self, "aliases"):
if not self.aliases: # type: ignore
aliases_str = "-- n/a --"
else:
aliases_str = ", ".join(self.aliases) # type: ignore
table.add_row("aliases", aliases_str)
table.add_row("kiara_id", str(self.kiara_id))
table.add_row("", "")
table.add_row("", Rule())
for k in sorted(self.__fields__.keys()):
if k in ["load_config", "value_id", "aliases", "kiara_id"]:
continue
attr = getattr(self, k)
if k in ["pedigree_output_name", "pedigree"]:
continue
elif k == "value_status":
v = f"[i]-- {attr.value} --[/i]"
else:
v = extract_renderable(attr)
table.add_row(k, v)
if (
show_pedigree
or show_load_config
or show_properties
or show_destinies
or show_destiny_backlinks
):
table.add_row("", "")
table.add_row("", Rule())
table.add_row("", "")
if show_pedigree:
pedigree = getattr(self, "pedigree")
if pedigree == ORPHAN:
v = "[i]-- external data --[/i]"
pedigree_output_name: Optional[Any] = None
else:
v = extract_renderable(pedigree)
pedigree_output_name = getattr(self, "pedigree_output_name")
row = ["pedigree", v]
table.add_row(*row)
if pedigree_output_name:
row = ["pedigree_output_name", pedigree_output_name]
table.add_row(*row)
if show_load_config:
load_config = self._data_registry.retrieve_load_config(self.value_id)
if load_config is None:
table.add_row("load_config", "-- no load config (yet?) --")
else:
table.add_row("load_config", load_config.create_renderable())
if show_properties:
if not self.property_links:
table.add_row("properties", "{}")
else:
properties = self._data_registry.load_values(self.property_links)
pr = properties.create_renderable(show_header=False)
table.add_row("properties", pr)
if hasattr(self, "destiny_links") and show_destinies:
if not self.destiny_links: # type: ignore
table.add_row("destinies", "{}")
else:
destinies = self._data_registry.load_values(self.destiny_links) # type: ignore
dr = destinies.create_renderable(show_header=False)
table.add_row("destinies", dr)
if show_destiny_backlinks:
if not self.destiny_backlinks:
table.add_row("destiny backlinks", "{}")
else:
destiny_items: List[Any] = []
for v_id, alias in self.destiny_backlinks.items():
destiny_items.append(Rule())
destiny_items.append(
f"[b]Value: [i]{v_id}[/i] (destiny alias: {alias})[/b]"
)
rendered = self._data_registry.render_data(
value_id=v_id, **render_config
)
destiny_items.append(rendered)
table.add_row("destiny backlinks", Group(*destiny_items))
if show_data:
rendered = self._data_registry.render_data(
self.value_id, target_type="terminal_renderable"
)
table.add_row("", "")
table.add_row("", Rule())
table.add_row("data preview", rendered)
return table
data: Any
property
readonly
¶data_type: DataType
property
readonly
¶destiny_backlinks: Mapping[uuid.UUID, str]
pydantic-field
¶Backlinks to values that this value acts as destiny/or property for.
is_initialized: bool
property
readonly
¶is_stored: bool
property
readonly
¶property_links: Mapping[str, uuid.UUID]
pydantic-field
¶Links to values that are properties of this value.
property_names: Iterable[str]
property
readonly
¶property_values: ValueMap
property
readonly
¶add_destiny_details(self, value_id, destiny_alias)
¶Source code in kiara/models/values/value.py
def add_destiny_details(self, value_id: uuid.UUID, destiny_alias: str):
if self._is_stored:
raise Exception(
f"Can't set destiny_refs to value '{self.value_id}': value already locked."
)
self.destiny_backlinks[value_id] = destiny_alias # type: ignore
add_property(self, value_id, property_path, add_origin_to_property_value=True)
¶Source code in kiara/models/values/value.py
def add_property(
self,
value_id: Union[uuid.UUID, "Value"],
property_path: str,
add_origin_to_property_value: bool = True,
):
value = None
try:
value_temp = value
value_id = value_id.value_id # type: ignore
value = value_temp
except Exception:
# in case a Value object was provided
pass
finally:
del value_temp
if add_origin_to_property_value:
if value is None:
value = self._data_registry.get_value(value_id=value_id) # type: ignore
if value._is_stored:
raise Exception(
f"Can't add property to value '{self.value_id}': referenced value '{value.value_id}' already locked, so it's not possible to add the property backlink (as requested)."
)
assert value is not None
if self._is_stored:
raise Exception(
f"Can't add property to value '{self.value_id}': value already locked."
)
if property_path in self.property_links.keys():
raise Exception(
f"Can't add property to value '{self.value_id}': property '{property_path}' already set."
)
self.property_links[property_path] = value_id # type: ignore
if add_origin_to_property_value:
value.add_destiny_details(
value_id=self.value_id, destiny_alias=property_path
)
self._cached_properties = None
create_renderable(self, **render_config)
¶Source code in kiara/models/values/value.py
def create_renderable(self, **render_config: Any) -> RenderableType:
from kiara.utils.output import extract_renderable
show_pedigree = render_config.get("show_pedigree", False)
show_load_config = render_config.get("show_load_config", False)
show_properties = render_config.get("show_properties", False)
show_destinies = render_config.get("show_destinies", False)
show_destiny_backlinks = render_config.get("show_destiny_backlinks", False)
show_data = render_config.get("show_data_preview", False)
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("Key", style="i")
table.add_column("Value")
table.add_row("value_id", str(self.value_id))
if hasattr(self, "aliases"):
if not self.aliases: # type: ignore
aliases_str = "-- n/a --"
else:
aliases_str = ", ".join(self.aliases) # type: ignore
table.add_row("aliases", aliases_str)
table.add_row("kiara_id", str(self.kiara_id))
table.add_row("", "")
table.add_row("", Rule())
for k in sorted(self.__fields__.keys()):
if k in ["load_config", "value_id", "aliases", "kiara_id"]:
continue
attr = getattr(self, k)
if k in ["pedigree_output_name", "pedigree"]:
continue
elif k == "value_status":
v = f"[i]-- {attr.value} --[/i]"
else:
v = extract_renderable(attr)
table.add_row(k, v)
if (
show_pedigree
or show_load_config
or show_properties
or show_destinies
or show_destiny_backlinks
):
table.add_row("", "")
table.add_row("", Rule())
table.add_row("", "")
if show_pedigree:
pedigree = getattr(self, "pedigree")
if pedigree == ORPHAN:
v = "[i]-- external data --[/i]"
pedigree_output_name: Optional[Any] = None
else:
v = extract_renderable(pedigree)
pedigree_output_name = getattr(self, "pedigree_output_name")
row = ["pedigree", v]
table.add_row(*row)
if pedigree_output_name:
row = ["pedigree_output_name", pedigree_output_name]
table.add_row(*row)
if show_load_config:
load_config = self._data_registry.retrieve_load_config(self.value_id)
if load_config is None:
table.add_row("load_config", "-- no load config (yet?) --")
else:
table.add_row("load_config", load_config.create_renderable())
if show_properties:
if not self.property_links:
table.add_row("properties", "{}")
else:
properties = self._data_registry.load_values(self.property_links)
pr = properties.create_renderable(show_header=False)
table.add_row("properties", pr)
if hasattr(self, "destiny_links") and show_destinies:
if not self.destiny_links: # type: ignore
table.add_row("destinies", "{}")
else:
destinies = self._data_registry.load_values(self.destiny_links) # type: ignore
dr = destinies.create_renderable(show_header=False)
table.add_row("destinies", dr)
if show_destiny_backlinks:
if not self.destiny_backlinks:
table.add_row("destiny backlinks", "{}")
else:
destiny_items: List[Any] = []
for v_id, alias in self.destiny_backlinks.items():
destiny_items.append(Rule())
destiny_items.append(
f"[b]Value: [i]{v_id}[/i] (destiny alias: {alias})[/b]"
)
rendered = self._data_registry.render_data(
value_id=v_id, **render_config
)
destiny_items.append(rendered)
table.add_row("destiny backlinks", Group(*destiny_items))
if show_data:
rendered = self._data_registry.render_data(
self.value_id, target_type="terminal_renderable"
)
table.add_row("", "")
table.add_row("", Rule())
table.add_row("data preview", rendered)
return table
get_property_data(self, property_key)
¶Source code in kiara/models/values/value.py
def get_property_data(self, property_key: str) -> Any:
return self.get_property_value(property_key=property_key).data
get_property_value(self, property_key)
¶Source code in kiara/models/values/value.py
def get_property_value(self, property_key) -> "Value":
if property_key not in self.property_links.keys():
raise Exception(
f"Value '{self.value_id}' has no property with key '{property_key}."
)
return self._data_registry.get_value(self.property_links[property_key])
retrieve_load_config(self)
¶Source code in kiara/models/values/value.py
def retrieve_load_config(self) -> Optional[LoadConfig]:
return self._data_registry.retrieve_load_config(value_id=self.value_id)
ValueDetails (KiaraModel)
pydantic-model
¶A wrapper class that manages and retieves value data and its details.
Source code in kiara/models/values/value.py
class ValueDetails(KiaraModel):
"""A wrapper class that manages and retieves value data and its details."""
value_id: uuid.UUID = Field(description="The id of the value.")
kiara_id: uuid.UUID = Field(
description="The id of the kiara context this value belongs to."
)
value_schema: ValueSchema = Field(
description="The schema that was used for this Value."
)
value_status: ValueStatus = Field(description="The set/unset status of this value.")
value_size: int = Field(description="The size of this value, in bytes.")
value_hash: int = Field(description="The hash of this value.")
pedigree: ValuePedigree = Field(
description="Information about the module and inputs that went into creating this value."
)
pedigree_output_name: str = Field(
description="The output name that produced this value (using the manifest inside the pedigree)."
)
data_type_class: PythonClass = Field(
description="The python class that is associtated with this model."
)
def _retrieve_id(self) -> str:
return str(self.value_id)
def _retrieve_category_id(self) -> str:
return VALUE_CATEGORY_ID
# @property
# def model_data_hash(self) -> int:
# return self._retrieve_data_to_hash()
def _retrieve_data_to_hash(self) -> Any:
return {"value_type": self.value_schema.type, "value_hash": self.value_hash}
@property
def data_type_name(self) -> str:
return self.value_schema.type
@property
def data_type_config(self) -> Mapping[str, Any]:
return self.value_schema.type_config
@property
def is_optional(self) -> bool:
return self.value_schema.optional
@property
def is_valid(self) -> bool:
"""Check whether the current value is valid"""
if self.is_optional:
return True
else:
return self.value_status == ValueStatus.SET
@property
def is_set(self) -> bool:
return self.value_status in [ValueStatus.SET, ValueStatus.DEFAULT]
@property
def value_status_string(self) -> str:
"""Print a human readable short description of this values status."""
if self.value_status == ValueStatus.DEFAULT:
return "set (default)"
elif self.value_status == ValueStatus.SET:
return "set"
elif self.value_status == ValueStatus.NONE:
result = "no value"
elif self.value_status == ValueStatus.NOT_SET:
result = "not set"
else:
raise Exception(
f"Invalid internal status of value '{self.value_id}'. This is most likely a bug."
)
if self.is_optional:
result = f"{result} (not required)"
return result
def __repr__(self):
return f"{self.__class__.__name__}(id={self.value_id}, type={self.data_type_name}, status={self.value_status.value})"
def __str__(self):
return self.__repr__()
data_type_class: PythonClass
pydantic-field
required
¶The python class that is associtated with this model.
data_type_config: Mapping[str, Any]
property
readonly
¶data_type_name: str
property
readonly
¶is_optional: bool
property
readonly
¶is_set: bool
property
readonly
¶is_valid: bool
property
readonly
¶Check whether the current value is valid
kiara_id: UUID
pydantic-field
required
¶The id of the kiara context this value belongs to.
pedigree: ValuePedigree
pydantic-field
required
¶Information about the module and inputs that went into creating this value.
pedigree_output_name: str
pydantic-field
required
¶The output name that produced this value (using the manifest inside the pedigree).
value_hash: int
pydantic-field
required
¶The hash of this value.
value_id: UUID
pydantic-field
required
¶The id of the value.
value_schema: ValueSchema
pydantic-field
required
¶The schema that was used for this Value.
value_size: int
pydantic-field
required
¶The size of this value, in bytes.
value_status: ValueStatus
pydantic-field
required
¶The set/unset status of this value.
value_status_string: str
property
readonly
¶Print a human readable short description of this values status.
ValueMap (KiaraModel, MutableMapping, Generic)
pydantic-model
¶Source code in kiara/models/values/value.py
class ValueMap(KiaraModel, MutableMapping[str, Value]): # type: ignore
values_schema: Dict[str, ValueSchema] = Field(
description="The schemas for all the values in this set."
)
@property
def field_names(self) -> Iterable[str]:
return sorted(self.values_schema.keys())
@abc.abstractmethod
def get_value_obj(self, field_name: str) -> Value:
pass
@property
def all_items_valid(self) -> bool:
for field_name in self.values_schema.keys():
item = self.get_value_obj(field_name)
if not item.is_valid:
return False
return True
def check_invalid(self) -> Dict[str, str]:
"""Check whether the value set is invalid, if it is, return a description of what's wrong."""
invalid: Dict[str, str] = {}
for field_name in self.values_schema.keys():
item = self.get_value_obj(field_name)
if not item.is_valid:
if item.value_schema.is_required():
if not item.is_set:
msg = "not set"
elif item.value_status == ValueStatus.NONE:
msg = "no value"
else:
msg = "n/a"
else:
msg = "n/a"
invalid[field_name] = msg
return invalid
def get_value_data_for_fields(
self, *field_names: str, raise_exception_when_unset: bool = False
) -> Dict[str, Any]:
"""Return the data for a one or several fields of this ValueMap.
If a value is unset, by default 'None' is returned for it. Unless 'raise_exception_when_unset' is set to 'True',
in which case an Exception will be raised (obviously).
"""
if raise_exception_when_unset:
unset: List[str] = []
for k in field_names:
v = self.get_value_obj(k)
if not v.is_set:
if raise_exception_when_unset:
unset.append(k)
if unset:
raise Exception(
f"Can't get data for fields, one or several of the requested fields are not set yet: {', '.join(unset)}."
)
result: Dict[str, Any] = {}
for k in field_names:
v = self.get_value_obj(k)
if not v.is_set:
result[k] = None
else:
result[k] = v.data
return result
def get_value_data(
self, field_name: str, raise_exception_when_unset: bool = False
) -> Any:
return self.get_value_data_for_fields(
field_name, raise_exception_when_unset=raise_exception_when_unset
)[field_name]
def get_all_value_ids(self) -> Dict[str, uuid.UUID]:
return {k: self.get_value_obj(k).value_id for k in self.field_names}
def get_all_value_data(
self, raise_exception_when_unset: bool = False
) -> Dict[str, Any]:
return self.get_value_data_for_fields(
*self.field_names,
raise_exception_when_unset=raise_exception_when_unset,
)
def set_values(self, **values) -> None:
for k, v in values.items():
self.set_value(k, v)
def set_value(self, field_name: str, data: Any) -> None:
raise Exception(
f"The value set implementation '{self.__class__.__name__}' is read-only, and does not support the setting or changing of values."
)
def __getitem__(self, item: str) -> Value:
return self.get_value_obj(item)
def __setitem__(self, key: str, value):
raise NotImplementedError()
# self.set_value(key, value)
def __delitem__(self, key: str):
raise Exception(f"Removing items not supported: {key}")
def __iter__(self):
return iter(self.field_names)
def __len__(self):
return len(list(self.values_schema))
def __repr__(self):
return f"{self.__class__.__name__}(field_names={self.field_names})"
def __str__(self):
return self.__repr__()
def create_invalid_renderable(self, **config) -> Optional[RenderableType]:
inv = self.check_invalid()
if not inv:
return None
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("field name", style="i")
table.add_column("details", style="b red")
for field, err in inv.items():
table.add_row(field, err)
return table
def create_renderable(self, **config: Any) -> RenderableType:
render_value_data = config.get("render_value_data", True)
field_title = config.get("field_title", "field")
value_title = config.get("value_title", "value")
show_header = config.get("show_header", True)
table = Table(show_lines=False, show_header=show_header, box=box.SIMPLE)
table.add_column(field_title, style="b")
table.add_column(value_title, style="i")
for field_name in self.field_names:
value = self.get_value_obj(field_name=field_name)
if render_value_data:
rendered = value._data_registry.render_data(
value_id=value.value_id, target_type="terminal_renderable", **config
)
else:
rendered = value.create_renderable(**config)
table.add_row(field_name, rendered)
return table
all_items_valid: bool
property
readonly
¶field_names: Iterable[str]
property
readonly
¶values_schema: Dict[str, kiara.models.values.value_schema.ValueSchema]
pydantic-field
required
¶The schemas for all the values in this set.
check_invalid(self)
¶Check whether the value set is invalid, if it is, return a description of what's wrong.
Source code in kiara/models/values/value.py
def check_invalid(self) -> Dict[str, str]:
"""Check whether the value set is invalid, if it is, return a description of what's wrong."""
invalid: Dict[str, str] = {}
for field_name in self.values_schema.keys():
item = self.get_value_obj(field_name)
if not item.is_valid:
if item.value_schema.is_required():
if not item.is_set:
msg = "not set"
elif item.value_status == ValueStatus.NONE:
msg = "no value"
else:
msg = "n/a"
else:
msg = "n/a"
invalid[field_name] = msg
return invalid
create_invalid_renderable(self, **config)
¶Source code in kiara/models/values/value.py
def create_invalid_renderable(self, **config) -> Optional[RenderableType]:
inv = self.check_invalid()
if not inv:
return None
table = Table(show_header=False, box=box.SIMPLE)
table.add_column("field name", style="i")
table.add_column("details", style="b red")
for field, err in inv.items():
table.add_row(field, err)
return table
create_renderable(self, **config)
¶Source code in kiara/models/values/value.py
def create_renderable(self, **config: Any) -> RenderableType:
render_value_data = config.get("render_value_data", True)
field_title = config.get("field_title", "field")
value_title = config.get("value_title", "value")
show_header = config.get("show_header", True)
table = Table(show_lines=False, show_header=show_header, box=box.SIMPLE)
table.add_column(field_title, style="b")
table.add_column(value_title, style="i")
for field_name in self.field_names:
value = self.get_value_obj(field_name=field_name)
if render_value_data:
rendered = value._data_registry.render_data(
value_id=value.value_id, target_type="terminal_renderable", **config
)
else:
rendered = value.create_renderable(**config)
table.add_row(field_name, rendered)
return table
get_all_value_data(self, raise_exception_when_unset=False)
¶Source code in kiara/models/values/value.py
def get_all_value_data(
self, raise_exception_when_unset: bool = False
) -> Dict[str, Any]:
return self.get_value_data_for_fields(
*self.field_names,
raise_exception_when_unset=raise_exception_when_unset,
)
get_all_value_ids(self)
¶Source code in kiara/models/values/value.py
def get_all_value_ids(self) -> Dict[str, uuid.UUID]:
return {k: self.get_value_obj(k).value_id for k in self.field_names}
get_value_data(self, field_name, raise_exception_when_unset=False)
¶Source code in kiara/models/values/value.py
def get_value_data(
self, field_name: str, raise_exception_when_unset: bool = False
) -> Any:
return self.get_value_data_for_fields(
field_name, raise_exception_when_unset=raise_exception_when_unset
)[field_name]
get_value_data_for_fields(self, *field_names, *, raise_exception_when_unset=False)
¶Return the data for a one or several fields of this ValueMap.
If a value is unset, by default 'None' is returned for it. Unless 'raise_exception_when_unset' is set to 'True', in which case an Exception will be raised (obviously).
Source code in kiara/models/values/value.py
def get_value_data_for_fields(
self, *field_names: str, raise_exception_when_unset: bool = False
) -> Dict[str, Any]:
"""Return the data for a one or several fields of this ValueMap.
If a value is unset, by default 'None' is returned for it. Unless 'raise_exception_when_unset' is set to 'True',
in which case an Exception will be raised (obviously).
"""
if raise_exception_when_unset:
unset: List[str] = []
for k in field_names:
v = self.get_value_obj(k)
if not v.is_set:
if raise_exception_when_unset:
unset.append(k)
if unset:
raise Exception(
f"Can't get data for fields, one or several of the requested fields are not set yet: {', '.join(unset)}."
)
result: Dict[str, Any] = {}
for k in field_names:
v = self.get_value_obj(k)
if not v.is_set:
result[k] = None
else:
result[k] = v.data
return result
get_value_obj(self, field_name)
¶Source code in kiara/models/values/value.py
@abc.abstractmethod
def get_value_obj(self, field_name: str) -> Value:
pass
set_value(self, field_name, data)
¶Source code in kiara/models/values/value.py
def set_value(self, field_name: str, data: Any) -> None:
raise Exception(
f"The value set implementation '{self.__class__.__name__}' is read-only, and does not support the setting or changing of values."
)
set_values(self, **values)
¶Source code in kiara/models/values/value.py
def set_values(self, **values) -> None:
for k, v in values.items():
self.set_value(k, v)
ValueMapReadOnly (ValueMap)
pydantic-model
¶Source code in kiara/models/values/value.py
class ValueMapReadOnly(ValueMap): # type: ignore
@classmethod
def create_from_ids(cls, data_registry: "DataRegistry", **value_ids: uuid.UUID):
values = {k: data_registry.get_value(v) for k, v in value_ids.items()}
return ValueMapReadOnly.construct(value_items=values)
value_items: Dict[str, Value] = Field(
description="The values contained in this set."
)
def _retrieve_id(self) -> str:
return str(uuid.uuid4())
def _retrieve_category_id(self) -> str:
return VALUES_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return {
k: self.get_value_obj(k).model_data_hash for k in self.values_schema.keys()
}
def get_value_obj(self, field_name: str) -> Value:
if field_name not in self.value_items.keys():
raise KeyError(
f"Field '{field_name}' not available in value set. Available fields: {', '.join(self.field_names)}"
)
return self.value_items[field_name]
value_items: Dict[str, kiara.models.values.value.Value]
pydantic-field
required
¶The values contained in this set.
create_from_ids(data_registry, **value_ids)
classmethod
¶Source code in kiara/models/values/value.py
@classmethod
def create_from_ids(cls, data_registry: "DataRegistry", **value_ids: uuid.UUID):
values = {k: data_registry.get_value(v) for k, v in value_ids.items()}
return ValueMapReadOnly.construct(value_items=values)
get_value_obj(self, field_name)
¶Source code in kiara/models/values/value.py
def get_value_obj(self, field_name: str) -> Value:
if field_name not in self.value_items.keys():
raise KeyError(
f"Field '{field_name}' not available in value set. Available fields: {', '.join(self.field_names)}"
)
return self.value_items[field_name]
ValueMapWritable (ValueMap)
pydantic-model
¶Source code in kiara/models/values/value.py
class ValueMapWritable(ValueMap): # type: ignore
@classmethod
def create_from_schema(
cls, kiara: "Kiara", schema: Mapping[str, ValueSchema], pedigree: ValuePedigree
) -> "ValueMapWritable":
v = ValueMapWritable(values_schema=schema, pedigree=pedigree)
v._data_registry = kiara.data_registry
return v
value_items: Dict[str, Value] = Field(
description="The values contained in this set.", default_factory=dict
)
pedigree: ValuePedigree = Field(
description="The pedigree to add to all of the result values."
)
_values_uncommitted: Dict[str, Any] = PrivateAttr(default_factory=dict)
_data_registry: "DataRegistry" = PrivateAttr(default=None)
_auto_commit: bool = PrivateAttr(default=True)
def _retrieve_id(self) -> str:
return str(uuid.uuid4())
def _retrieve_category_id(self) -> str:
return VALUES_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return {
k: self.get_value_obj(k).model_data_hash for k in self.values_schema.keys()
}
def get_value_obj(self, field_name: str) -> Value:
"""Retrieve the value object for the specified field.
This class only creates the actual value object the first time it is requested, because there is a potential
cost to assembling it, and it might not be needed ever.
"""
if field_name not in self.values_schema.keys():
raise Exception(
f"Can't set data for field '{field_name}': field not valid, valid field names: {', '.join(self.field_names)}."
)
if field_name in self.value_items.keys():
return self.value_items[field_name]
elif field_name not in self._values_uncommitted.keys():
raise Exception(
f"Can't retrieve value for field '{field_name}': value not set (yet)."
)
schema = self.values_schema[field_name]
value_data = self._values_uncommitted[field_name]
if isinstance(value_data, Value):
value = value_data
elif isinstance(value_data, uuid.UUID):
value = self._data_registry.get_value(value_data)
else:
value = self._data_registry.register_data(
data=value_data,
schema=schema,
pedigree=self.pedigree,
pedigree_output_name=field_name,
reuse_existing=False,
)
self._values_uncommitted.pop(field_name)
self.value_items[field_name] = value
return self.value_items[field_name]
def sync_values(self):
for field_name in self.field_names:
self.get_value_obj(field_name)
invalid = self.check_invalid()
if invalid:
if is_debug():
import traceback
traceback.print_stack()
raise InvalidValuesException(invalid_values=invalid)
def set_value(self, field_name: str, data: Any) -> None:
"""Set the value for the specified field."""
if field_name not in self.field_names:
raise Exception(
f"Can't set data for field '{field_name}': field not valid, valid field names: {', '.join(self.field_names)}."
)
if self.value_items.get(field_name, False):
raise Exception(
f"Can't set data for field '{field_name}': field already committed."
)
if self._values_uncommitted.get(field_name, None) is not None:
raise Exception(
f"Can't set data for field '{field_name}': field already set."
)
self._values_uncommitted[field_name] = data
if self._auto_commit:
self.get_value_obj(field_name=field_name)
pedigree: ValuePedigree
pydantic-field
required
¶The pedigree to add to all of the result values.
value_items: Dict[str, kiara.models.values.value.Value]
pydantic-field
¶The values contained in this set.
create_from_schema(kiara, schema, pedigree)
classmethod
¶Source code in kiara/models/values/value.py
@classmethod
def create_from_schema(
cls, kiara: "Kiara", schema: Mapping[str, ValueSchema], pedigree: ValuePedigree
) -> "ValueMapWritable":
v = ValueMapWritable(values_schema=schema, pedigree=pedigree)
v._data_registry = kiara.data_registry
return v
get_value_obj(self, field_name)
¶Retrieve the value object for the specified field.
This class only creates the actual value object the first time it is requested, because there is a potential cost to assembling it, and it might not be needed ever.
Source code in kiara/models/values/value.py
def get_value_obj(self, field_name: str) -> Value:
"""Retrieve the value object for the specified field.
This class only creates the actual value object the first time it is requested, because there is a potential
cost to assembling it, and it might not be needed ever.
"""
if field_name not in self.values_schema.keys():
raise Exception(
f"Can't set data for field '{field_name}': field not valid, valid field names: {', '.join(self.field_names)}."
)
if field_name in self.value_items.keys():
return self.value_items[field_name]
elif field_name not in self._values_uncommitted.keys():
raise Exception(
f"Can't retrieve value for field '{field_name}': value not set (yet)."
)
schema = self.values_schema[field_name]
value_data = self._values_uncommitted[field_name]
if isinstance(value_data, Value):
value = value_data
elif isinstance(value_data, uuid.UUID):
value = self._data_registry.get_value(value_data)
else:
value = self._data_registry.register_data(
data=value_data,
schema=schema,
pedigree=self.pedigree,
pedigree_output_name=field_name,
reuse_existing=False,
)
self._values_uncommitted.pop(field_name)
self.value_items[field_name] = value
return self.value_items[field_name]
set_value(self, field_name, data)
¶Set the value for the specified field.
Source code in kiara/models/values/value.py
def set_value(self, field_name: str, data: Any) -> None:
"""Set the value for the specified field."""
if field_name not in self.field_names:
raise Exception(
f"Can't set data for field '{field_name}': field not valid, valid field names: {', '.join(self.field_names)}."
)
if self.value_items.get(field_name, False):
raise Exception(
f"Can't set data for field '{field_name}': field already committed."
)
if self._values_uncommitted.get(field_name, None) is not None:
raise Exception(
f"Can't set data for field '{field_name}': field already set."
)
self._values_uncommitted[field_name] = data
if self._auto_commit:
self.get_value_obj(field_name=field_name)
sync_values(self)
¶Source code in kiara/models/values/value.py
def sync_values(self):
for field_name in self.field_names:
self.get_value_obj(field_name)
invalid = self.check_invalid()
if invalid:
if is_debug():
import traceback
traceback.print_stack()
raise InvalidValuesException(invalid_values=invalid)
ValuePedigree (InputsManifest)
pydantic-model
¶Source code in kiara/models/values/value.py
class ValuePedigree(InputsManifest):
kiara_id: uuid.UUID = Field(
description="The id of the kiara context a value was created in."
)
environments: Dict[str, int] = Field(
description="References to the runtime environment details a value was created in."
)
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return VALUE_PEDIGREE_TYPE_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return {
"manifest": self.manifest_hash,
"inputs": self.inputs_hash,
# "environments": self.environments
}
def __repr__(self):
return f"ValuePedigree(module_type={self.module_type}, inputs=[{', '.join(self.inputs.keys())}], hash={self.model_data_hash})"
def __str__(self):
return self.__repr__()
value_metadata
special
¶
MetadataTypeClassesInfo (TypeInfoModelGroup)
pydantic-model
¶Source code in kiara/models/values/value_metadata/__init__.py
class MetadataTypeClassesInfo(TypeInfoModelGroup):
@classmethod
def base_info_class(cls) -> Type[KiaraTypeInfoModel]:
return MetadataTypeInfoModel
type_name: Literal["value_metadata"] = "value_metadata"
type_infos: Mapping[str, MetadataTypeInfoModel] = Field(
description="The value metadata info instances for each type."
)
type_infos: Mapping[str, kiara.models.values.value_metadata.MetadataTypeInfoModel]
pydantic-field
required
¶The value metadata info instances for each type.
type_name: Literal['value_metadata']
pydantic-field
¶base_info_class()
classmethod
¶Source code in kiara/models/values/value_metadata/__init__.py
@classmethod
def base_info_class(cls) -> Type[KiaraTypeInfoModel]:
return MetadataTypeInfoModel
MetadataTypeInfoModel (KiaraTypeInfoModel)
pydantic-model
¶Source code in kiara/models/values/value_metadata/__init__.py
class MetadataTypeInfoModel(KiaraTypeInfoModel):
@classmethod
def create_from_type_class(
self, type_cls: Type[ValueMetadata]
) -> "MetadataTypeInfoModel":
authors_md = AuthorsMetadataModel.from_class(type_cls)
doc = DocumentationMetadataModel.from_class_doc(type_cls)
python_class = PythonClass.from_class(type_cls)
properties_md = ContextMetadataModel.from_class(type_cls)
type_name = type_cls._metadata_key # type: ignore
schema = type_cls.schema()
return MetadataTypeInfoModel.construct(
type_name=type_name,
documentation=doc,
authors=authors_md,
context=properties_md,
python_class=python_class,
metadata_schema=schema,
)
@classmethod
def base_class(self) -> Type[ValueMetadata]:
return ValueMetadata
@classmethod
def category_name(cls) -> str:
return "value_metadata"
metadata_schema: Dict[str, Any] = Field(
description="The (json) schema for this metadata value."
)
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
include_schema = config.get("include_schema", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
if hasattr(self, "python_class"):
table.add_row("Python class", self.python_class.create_renderable())
if include_schema:
schema = Syntax(
orjson_dumps(self.metadata_schema, option=orjson.OPT_INDENT_2),
"json",
background_color="default",
)
table.add_row("metadata_schema", schema)
return table
metadata_schema: Dict[str, Any]
pydantic-field
required
¶The (json) schema for this metadata value.
base_class()
classmethod
¶Source code in kiara/models/values/value_metadata/__init__.py
@classmethod
def base_class(self) -> Type[ValueMetadata]:
return ValueMetadata
category_name()
classmethod
¶Source code in kiara/models/values/value_metadata/__init__.py
@classmethod
def category_name(cls) -> str:
return "value_metadata"
create_from_type_class(type_cls)
classmethod
¶Source code in kiara/models/values/value_metadata/__init__.py
@classmethod
def create_from_type_class(
self, type_cls: Type[ValueMetadata]
) -> "MetadataTypeInfoModel":
authors_md = AuthorsMetadataModel.from_class(type_cls)
doc = DocumentationMetadataModel.from_class_doc(type_cls)
python_class = PythonClass.from_class(type_cls)
properties_md = ContextMetadataModel.from_class(type_cls)
type_name = type_cls._metadata_key # type: ignore
schema = type_cls.schema()
return MetadataTypeInfoModel.construct(
type_name=type_name,
documentation=doc,
authors=authors_md,
context=properties_md,
python_class=python_class,
metadata_schema=schema,
)
create_renderable(self, **config)
¶Source code in kiara/models/values/value_metadata/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
include_doc = config.get("include_doc", True)
include_schema = config.get("include_schema", True)
table = Table(box=box.SIMPLE, show_header=False, padding=(0, 0, 0, 0))
table.add_column("property", style="i")
table.add_column("value")
if include_doc:
table.add_row(
"Documentation",
Panel(self.documentation.create_renderable(), box=box.SIMPLE),
)
table.add_row("Author(s)", self.authors.create_renderable())
table.add_row("Context", self.context.create_renderable())
if hasattr(self, "python_class"):
table.add_row("Python class", self.python_class.create_renderable())
if include_schema:
schema = Syntax(
orjson_dumps(self.metadata_schema, option=orjson.OPT_INDENT_2),
"json",
background_color="default",
)
table.add_row("metadata_schema", schema)
return table
ValueMetadata (KiaraModel)
pydantic-model
¶Source code in kiara/models/values/value_metadata/__init__.py
class ValueMetadata(KiaraModel):
@classmethod
@abc.abstractmethod
def retrieve_supported_data_types(cls) -> Iterable[str]:
pass
@classmethod
@abc.abstractmethod
def create_value_metadata(
cls, value: "Value"
) -> Union["ValueMetadata", Dict[str, Any]]:
pass
# @property
# def metadata_key(self) -> str:
# return self._metadata_key # type: ignore # this is added by the kiara class loading functionality
def _retrieve_id(self) -> str:
return self._metadata_key # type: ignore
def _retrieve_category_id(self) -> str:
return f"{VALUE_METADATA_CATEGORY_ID}.{self._metadata_key}" # type: ignore
def _retrieve_data_to_hash(self) -> Any:
return {"metadata": self.dict(), "schema": self.schema_json()}
create_value_metadata(value)
classmethod
¶Source code in kiara/models/values/value_metadata/__init__.py
@classmethod
@abc.abstractmethod
def create_value_metadata(
cls, value: "Value"
) -> Union["ValueMetadata", Dict[str, Any]]:
pass
retrieve_supported_data_types()
classmethod
¶Source code in kiara/models/values/value_metadata/__init__.py
@classmethod
@abc.abstractmethod
def retrieve_supported_data_types(cls) -> Iterable[str]:
pass
included_metadata_types
special
¶
FileBundleMetadata (ValueMetadata)
pydantic-model
¶File stats.
Source code in kiara/models/values/value_metadata/included_metadata_types/__init__.py
class FileBundleMetadata(ValueMetadata):
"""File stats."""
_metadata_key = "file_bundle"
@classmethod
def retrieve_supported_data_types(cls) -> Iterable[str]:
return ["file_bundle"]
@classmethod
def create_value_metadata(cls, value: "Value") -> "FileBundleMetadata":
return FileBundleMetadata.construct(file_bundle=value.data)
file_bundle: FileBundle = Field(description="The file-specific metadata.")
file_bundle: FileBundle
pydantic-field
required
¶The file-specific metadata.
create_value_metadata(value)
classmethod
¶Source code in kiara/models/values/value_metadata/included_metadata_types/__init__.py
@classmethod
def create_value_metadata(cls, value: "Value") -> "FileBundleMetadata":
return FileBundleMetadata.construct(file_bundle=value.data)
retrieve_supported_data_types()
classmethod
¶Source code in kiara/models/values/value_metadata/included_metadata_types/__init__.py
@classmethod
def retrieve_supported_data_types(cls) -> Iterable[str]:
return ["file_bundle"]
FileMetadata (ValueMetadata)
pydantic-model
¶File stats.
Source code in kiara/models/values/value_metadata/included_metadata_types/__init__.py
class FileMetadata(ValueMetadata):
"""File stats."""
_metadata_key = "file"
@classmethod
def retrieve_supported_data_types(cls) -> Iterable[str]:
return ["file"]
@classmethod
def create_value_metadata(cls, value: "Value") -> "FileMetadata":
return FileMetadata.construct(file=value.data)
file: FileModel = Field(description="The file-specific metadata.")
file: FileModel
pydantic-field
required
¶The file-specific metadata.
create_value_metadata(value)
classmethod
¶Source code in kiara/models/values/value_metadata/included_metadata_types/__init__.py
@classmethod
def create_value_metadata(cls, value: "Value") -> "FileMetadata":
return FileMetadata.construct(file=value.data)
retrieve_supported_data_types()
classmethod
¶Source code in kiara/models/values/value_metadata/included_metadata_types/__init__.py
@classmethod
def retrieve_supported_data_types(cls) -> Iterable[str]:
return ["file"]
PythonClassMetadata (ValueMetadata)
pydantic-model
¶Python class and module information.
Source code in kiara/models/values/value_metadata/included_metadata_types/__init__.py
class PythonClassMetadata(ValueMetadata):
"""Python class and module information."""
_metadata_key = "python_class"
@classmethod
def retrieve_supported_data_types(cls) -> Iterable[str]:
return ["any"]
@classmethod
def create_value_metadata(cls, value: "Value") -> "PythonClassMetadata":
return PythonClassMetadata.construct(
python_class=PythonClass.from_class(value.data.__class__)
)
# metadata_key: Literal["python_class"]
python_class: PythonClass = Field(
description="Details about the Python class that backs this value."
)
python_class: PythonClass
pydantic-field
required
¶Details about the Python class that backs this value.
create_value_metadata(value)
classmethod
¶Source code in kiara/models/values/value_metadata/included_metadata_types/__init__.py
@classmethod
def create_value_metadata(cls, value: "Value") -> "PythonClassMetadata":
return PythonClassMetadata.construct(
python_class=PythonClass.from_class(value.data.__class__)
)
retrieve_supported_data_types()
classmethod
¶Source code in kiara/models/values/value_metadata/included_metadata_types/__init__.py
@classmethod
def retrieve_supported_data_types(cls) -> Iterable[str]:
return ["any"]
value_schema
¶
ValueSchema (KiaraModel)
pydantic-model
¶The schema of a value.
The schema contains the [ValueTypeOrm][kiara.data.values.ValueTypeOrm] of a value, as well as an optional default that will be used if no user input was given (yet) for a value.
For more complex container data_types like array, tables, unions etc, data_types can also be configured with values from the type_config field.
Source code in kiara/models/values/value_schema.py
class ValueSchema(KiaraModel):
"""The schema of a value.
The schema contains the [ValueTypeOrm][kiara.data.values.ValueTypeOrm] of a value, as well as an optional default that
will be used if no user input was given (yet) for a value.
For more complex container data_types like array, tables, unions etc, data_types can also be configured with values from the ``type_config`` field.
"""
class Config:
use_enum_values = True
# extra = Extra.forbid
type: str = Field(description="The type of the value.")
type_config: typing.Dict[str, typing.Any] = Field(
description="Configuration for the type, in case it's complex.",
default_factory=dict,
)
default: typing.Any = Field(
description="A default value.", default=SpecialValue.NOT_SET
)
optional: bool = Field(
description="Whether this value is required (True), or whether 'None' value is allowed (False).",
default=False,
)
is_constant: bool = Field(
description="Whether the value is a constant.", default=False
)
doc: DocumentationMetadataModel = Field(
default="-- n/a --",
description="A description for the value of this input field.",
)
@validator("doc", pre=True)
def validate_doc(cls, value):
return DocumentationMetadataModel.create(value)
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return VALUE_SCHEMA_CATEGORY_ID
def _retrieve_data_to_hash(self) -> typing.Any:
return {"type": self.type, "type_config": self.type_config}
def is_required(self):
if self.optional:
return False
else:
if self.default in [None, SpecialValue.NOT_SET, SpecialValue.NO_VALUE]:
return True
else:
return False
# def validate_types(self, kiara: "Kiara"):
#
# if self.type not in kiara.value_type_names:
# raise ValueError(
# f"Invalid value type '{self.type}', available data_types: {kiara.value_type_names}"
# )
def __eq__(self, other):
if not isinstance(other, ValueSchema):
return False
return (self.type, self.default) == (other.type, other.default)
def __hash__(self):
return hash((self.type, self.default))
def __repr__(self):
return f"ValueSchema(type={self.type}, default={self.default}, optional={self.optional})"
def __str__(self):
return self.__repr__()
default: Any
pydantic-field
¶A default value.
doc: DocumentationMetadataModel
pydantic-field
¶A description for the value of this input field.
is_constant: bool
pydantic-field
¶Whether the value is a constant.
optional: bool
pydantic-field
¶Whether this value is required (True), or whether 'None' value is allowed (False).
type: str
pydantic-field
required
¶The type of the value.
type_config: Dict[str, Any]
pydantic-field
¶Configuration for the type, in case it's complex.
Config
¶Source code in kiara/models/values/value_schema.py
class Config:
use_enum_values = True
# extra = Extra.forbid
is_required(self)
¶Source code in kiara/models/values/value_schema.py
def is_required(self):
if self.optional:
return False
else:
if self.default in [None, SpecialValue.NOT_SET, SpecialValue.NO_VALUE]:
return True
else:
return False
validate_doc(value)
classmethod
¶Source code in kiara/models/values/value_schema.py
@validator("doc", pre=True)
def validate_doc(cls, value):
return DocumentationMetadataModel.create(value)
modules
special
¶
KIARA_CONFIG
¶
ValueSetSchema
¶
yaml
¶
Classes¶
InputOutputObject (ABC)
¶
Abstract base class for classes that define inputs and outputs schemas.
Both the 'create_inputs_schemaandcreawte_outputs_schema` methods implemented by child classes return a description of the input schema of this module.
If returning a dictionary of dictionaries, the format of the return value is as follows (items with '*' are optional):
{
"[input_field_name]: {
"type": "[type]",
"doc*": "[a description of this input]",
"optional*': [boolean whether this input is optional or required (defaults to 'False')]
"[other_input_field_name]: {
"type: ...
...
}
Source code in kiara/modules/__init__.py
class InputOutputObject(abc.ABC):
"""Abstract base class for classes that define inputs and outputs schemas.
Both the 'create_inputs_schema` and `creawte_outputs_schema` methods implemented by child classes return a description of the input schema of this module.
If returning a dictionary of dictionaries, the format of the return value is as follows (items with '*' are optional):
```
{
"[input_field_name]: {
"type": "[type]",
"doc*": "[a description of this input]",
"optional*': [boolean whether this input is optional or required (defaults to 'False')]
"[other_input_field_name]: {
"type: ...
...
}
```
"""
def __init__(
self,
alias: str,
config: KiaraModuleConfig = None,
allow_empty_inputs_schema: bool = False,
allow_empty_outputs_schema: bool = False,
):
self._alias: str = alias
self._inputs_schema: Mapping[str, ValueSchema] = None # type: ignore
self._outputs_schema: Mapping[str, ValueSchema] = None # type: ignore
self._constants: Mapping[str, ValueSchema] = None # type: ignore
if config is None:
config = KiaraModuleConfig()
self._config: KiaraModuleConfig = config
self._allow_empty_inputs: bool = allow_empty_inputs_schema
self._allow_empty_outputs: bool = allow_empty_outputs_schema
@property
def alias(self) -> str:
return self._alias
def input_required(self, input_name: str):
if input_name not in self._inputs_schema.keys():
raise Exception(
f"No input '{input_name}', available inputs: {', '.join(self._inputs_schema)}"
)
if not self._inputs_schema[input_name].is_required():
return False
if input_name in self.constants.keys():
return False
else:
return True
@abstractmethod
def create_inputs_schema(
self,
) -> ValueSetSchema:
"""Return the schema for this types' inputs."""
@abstractmethod
def create_outputs_schema(
self,
) -> ValueSetSchema:
"""Return the schema for this types' outputs."""
@property
def inputs_schema(self) -> Mapping[str, ValueSchema]:
"""The input schema for this module."""
if self._inputs_schema is None:
self._create_inputs_schema()
return self._inputs_schema # type: ignore
@property
def constants(self) -> Mapping[str, ValueSchema]:
if self._constants is None:
self._create_inputs_schema()
return self._constants # type: ignore
def _create_inputs_schema(self) -> None:
try:
_input_schemas_data = self.create_inputs_schema()
if _input_schemas_data is None:
raise Exception(
f"Invalid inputs implementation for '{self.alias}': no inputs schema"
)
if not _input_schemas_data and not self._allow_empty_inputs:
raise Exception(
f"Invalid inputs implementation for '{self.alias}': empty inputs schema"
)
try:
_input_schemas = create_schema_dict(schema_config=_input_schemas_data)
except Exception as e:
raise Exception(f"Can't create input schemas for '{self.alias}': {e}")
defaults = self._config.defaults
constants = self._config.constants
for k, v in defaults.items():
if k not in _input_schemas.keys():
raise Exception(
f"Can't create inputs for '{self.alias}', invalid default field name '{k}'. Available field names: '{', '.join(_input_schemas.keys())}'" # type: ignore
)
for k, v in constants.items():
if k not in _input_schemas.keys():
raise Exception(
f"Can't create inputs for '{self.alias}', invalid constant field name '{k}'. Available field names: '{', '.join(_input_schemas.keys())}'" # type: ignore
)
self._inputs_schema, self._constants = overlay_constants_and_defaults(
_input_schemas, defaults=defaults, constants=constants
)
except Exception as e:
raise Exception(f"Can't create input schemas for instance '{self.alias}': {e}") # type: ignore
@property
def outputs_schema(self) -> Mapping[str, ValueSchema]:
"""The output schema for this module."""
if self._outputs_schema is not None:
return self._outputs_schema
try:
_output_schema = self.create_outputs_schema()
if _output_schema is None:
raise Exception(
f"Invalid outputs implementation for '{self.alias}': no outputs schema"
)
if not _output_schema and not self._allow_empty_outputs:
raise Exception(
f"Invalid outputs implementation for '{self.alias}': empty outputs schema"
)
try:
self._outputs_schema = create_schema_dict(schema_config=_output_schema)
except Exception as e:
raise Exception(
f"Can't create output schemas for module {self.alias}: {e}"
)
return self._outputs_schema
except Exception as e:
raise Exception(f"Can't create output schemas for instance '{self.alias}': {e}") # type: ignore
@property
def input_names(self) -> Iterable[str]:
"""A list of input field names for this module."""
return self.inputs_schema.keys()
@property
def output_names(self) -> Iterable[str]:
"""A list of output field names for this module."""
return self.outputs_schema.keys()
def augment_module_inputs(self, inputs: Mapping[str, Any]) -> Dict[str, Any]:
return augment_values(
values=inputs, schemas=self.inputs_schema, constants=self.constants
)
# def augment_outputs(self, outputs: Mapping[str, Any]) -> Dict[str, Any]:
# return augment_values(values=outputs, schemas=self.outputs_schema)
Attributes¶
alias: str
property
readonly
¶constants: Mapping[str, kiara.models.values.value_schema.ValueSchema]
property
readonly
¶input_names: Iterable[str]
property
readonly
¶A list of input field names for this module.
inputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
property
readonly
¶The input schema for this module.
output_names: Iterable[str]
property
readonly
¶A list of output field names for this module.
outputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
property
readonly
¶The output schema for this module.
Methods¶
augment_module_inputs(self, inputs)
¶Source code in kiara/modules/__init__.py
def augment_module_inputs(self, inputs: Mapping[str, Any]) -> Dict[str, Any]:
return augment_values(
values=inputs, schemas=self.inputs_schema, constants=self.constants
)
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/__init__.py
@abstractmethod
def create_inputs_schema(
self,
) -> ValueSetSchema:
"""Return the schema for this types' inputs."""
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/__init__.py
@abstractmethod
def create_outputs_schema(
self,
) -> ValueSetSchema:
"""Return the schema for this types' outputs."""
input_required(self, input_name)
¶Source code in kiara/modules/__init__.py
def input_required(self, input_name: str):
if input_name not in self._inputs_schema.keys():
raise Exception(
f"No input '{input_name}', available inputs: {', '.join(self._inputs_schema)}"
)
if not self._inputs_schema[input_name].is_required():
return False
if input_name in self.constants.keys():
return False
else:
return True
KiaraModule (InputOutputObject, Generic)
¶
The base class that every custom module in Kiara needs to inherit from.
The core of every KiaraModule is a process method, which should be a 'pure',
idempotent function that creates one or several output values from the given input(s), and its purpose is to transfor
a set of inputs into a set of outputs.
Every module can be configured. The module configuration schema can differ, but every one such configuration needs to
subclass the [KiaraModuleConfig][kiara.module_config.KiaraModuleConfig] class and set as the value to the
_config_cls attribute of the module class. This is useful, because it allows for some modules to serve a much
larger variety of use-cases than non-configurable modules would be, which would mean more code duplication because
of very simlilar, but slightly different module data_types.
Each module class (type) has a unique -- within a kiara context -- module type id which can be accessed via the
_module_type_name class attribute.
Examples:
A simple example would be an 'addition' module, with a and b configured as inputs, and z as the output field name.
An implementing class would look something like this:
TODO
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
module_config |
Union[NoneType, ~KIARA_CONFIG, Mapping[str, Any]] |
the configuation for this module |
None |
Source code in kiara/modules/__init__.py
class KiaraModule(InputOutputObject, Generic[KIARA_CONFIG]):
"""The base class that every custom module in *Kiara* needs to inherit from.
The core of every ``KiaraModule`` is a ``process`` method, which should be a 'pure',
idempotent function that creates one or several output values from the given input(s), and its purpose is to transfor
a set of inputs into a set of outputs.
Every module can be configured. The module configuration schema can differ, but every one such configuration needs to
subclass the [KiaraModuleConfig][kiara.module_config.KiaraModuleConfig] class and set as the value to the
``_config_cls`` attribute of the module class. This is useful, because it allows for some modules to serve a much
larger variety of use-cases than non-configurable modules would be, which would mean more code duplication because
of very simlilar, but slightly different module data_types.
Each module class (type) has a unique -- within a *kiara* context -- module type id which can be accessed via the
``_module_type_name`` class attribute.
Examples:
A simple example would be an 'addition' module, with ``a`` and ``b`` configured as inputs, and ``z`` as the output field name.
An implementing class would look something like this:
TODO
Arguments:
module_config: the configuation for this module
"""
# TODO: not quite sure about this generic type here, mypy doesn't seem to like it
_config_cls: Type[KIARA_CONFIG] = KiaraModuleConfig # type: ignore
@classmethod
def is_pipeline(cls) -> bool:
"""Check whether this module type is a pipeline, or not."""
return False
@classmethod
def _calculate_module_hash(
cls, module_type_config: Union[Mapping[str, Any], KIARA_CONFIG]
):
if isinstance(module_type_config, Mapping):
module_type_config = cls._config_cls(**module_type_config)
obj = {
"module_type": cls._module_type_name, # type: ignore
"module_type_config": module_type_config.model_data_hash, # type: ignore
}
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
return h[obj]
def __init__(
self,
module_config: Union[None, KIARA_CONFIG, Mapping[str, Any]] = None,
):
self._id: uuid.UUID = uuid.uuid4()
if isinstance(module_config, KiaraModuleConfig):
self._config: KIARA_CONFIG = module_config # type: ignore
elif module_config is None:
self._config = self.__class__._config_cls()
elif isinstance(module_config, Mapping):
try:
self._config = self.__class__._config_cls(**module_config)
except ValidationError as ve:
raise KiaraModuleConfigException(
f"Error creating module '{id}'. {ve}",
self.__class__,
module_config,
ve,
)
else:
raise TypeError(f"Invalid type for module config: {type(module_config)}")
self._module_hash: Optional[int] = None
self._characteristics: Optional[ModuleCharacteristics] = None
super().__init__(alias=self.__class__._module_type_name, config=self._config) # type: ignore
self._operation: Optional[Operation] = None
# self._merged_input_schemas: typing.Mapping[str, ValueSchema] = None # type: ignore
@property
def module_id(self) -> uuid.UUID:
"""The id of this module."""
return self._id
@property
def module_type_name(self) -> str:
return self._module_type_name # type: ignore
@property
def config(self) -> KIARA_CONFIG:
"""Retrieve the configuration object for this module.
Returns:
the module-class-specific config object
"""
return self._config
@property
def module_instance_hash(self) -> int:
if self._module_hash is None:
self._module_hash = self.__class__._calculate_module_hash(self._config)
return self._module_hash
@property
def characteristics(self) -> ModuleCharacteristics:
if self._characteristics is not None:
return self._characteristics
self._characteristics = self._retrieve_module_characteristics()
return self._characteristics
def _retrieve_module_characteristics(self) -> ModuleCharacteristics:
return ModuleCharacteristics()
def get_config_value(self, key: str) -> Any:
"""Retrieve the value for a specific configuration option.
Arguments:
key: the config key
Returns:
the value for the provided key
"""
try:
return self.config.get(key)
except Exception:
raise Exception(
f"Error accessing config value '{key}' in module {self.__class__._module_type_name}." # type: ignore
)
def process_step(
self, inputs: "ValueMap", outputs: "ValueMap", job_log: JobLog
) -> None:
"""Kick off processing for a specific set of input/outputs.
This method calls the implemented [process][kiara.module.KiaraModule.process] method of the inheriting class,
as well as wrapping input/output-data related functionality.
Arguments:
inputs: the input value set
outputs: the output value set
"""
signature = inspect.signature(self.process) # type: ignore
if "job_log" not in signature.parameters.keys():
try:
self.process(inputs=inputs, outputs=outputs) # type: ignore
except Exception as e:
if is_debug():
try:
import traceback
traceback.print_exc()
except Exception:
pass
raise e
else:
try:
self.process(inputs=inputs, outputs=outputs, job_log=job_log) # type: ignore
except Exception as e:
if is_debug():
try:
import traceback
traceback.print_exc()
except Exception:
pass
raise e
def __eq__(self, other):
if self.__class__ != other.__class__:
return False
return self.module_instance_hash == other.module_instance_hash
def __hash__(self):
return self.module_instance_hash
def __repr__(self):
return f"{self.__class__.__name__}(id={self.module_id} module_type={self.module_type_name} input_names={list(self.input_names)} output_names={list(self.output_names)})"
def create_renderable(self, **config) -> RenderableType:
if self._operation is not None:
return self._operation
from kiara.models.module.operation import Operation
self._operation = Operation.create_from_module(self)
return self._operation
Attributes¶
characteristics: ModuleCharacteristics
property
readonly
¶config: ~KIARA_CONFIG
property
readonly
¶Retrieve the configuration object for this module.
Returns:
| Type | Description |
|---|---|
~KIARA_CONFIG |
the module-class-specific config object |
module_id: UUID
property
readonly
¶The id of this module.
module_instance_hash: int
property
readonly
¶module_type_name: str
property
readonly
¶Classes¶
_config_cls (KiaraModel)
private
pydantic-model
¶Base class that describes the configuration a [KiaraModule][kiara.module.KiaraModule] class accepts.
This is stored in the _config_cls class attribute in each KiaraModule class.
There are two config options every KiaraModule supports:
constants, anddefaults
Constants are pre-set inputs, and users can't change them and an error is thrown if they try. Defaults are default values that override the schema defaults, and those can be overwritten by users. If both a constant and a default value is set for an input field, an error is thrown.
Source code in kiara/modules/__init__.py
class KiaraModuleConfig(KiaraModel):
"""Base class that describes the configuration a [``KiaraModule``][kiara.module.KiaraModule] class accepts.
This is stored in the ``_config_cls`` class attribute in each ``KiaraModule`` class.
There are two config options every ``KiaraModule`` supports:
- ``constants``, and
- ``defaults``
Constants are pre-set inputs, and users can't change them and an error is thrown if they try. Defaults are default
values that override the schema defaults, and those can be overwritten by users. If both a constant and a default
value is set for an input field, an error is thrown.
"""
@classmethod
def requires_config(cls, config: Optional[Mapping[str, Any]] = None) -> bool:
"""Return whether this class can be used as-is, or requires configuration before an instance can be created."""
for field_name, field in cls.__fields__.items():
if field.required and field.default is None:
if config:
if config.get(field_name, None) is None:
return True
else:
return True
return False
_config_hash: str = PrivateAttr(default=None)
constants: Dict[str, Any] = Field(
default_factory=dict, description="Value constants for this module."
)
defaults: Dict[str, Any] = Field(
default_factory=dict, description="Value defaults for this module."
)
class Config:
extra = Extra.forbid
validate_assignment = True
def get(self, key: str) -> Any:
"""Get the value for the specified configuation key."""
if key not in self.__fields__:
raise Exception(
f"No config value '{key}' in module config class '{self.__class__.__name__}'."
)
return getattr(self, key)
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return MODULE_CONFIG_SCHEMA_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
def create_renderable(self, **config: Any) -> RenderableType:
my_table = Table(box=box.MINIMAL, show_header=False)
my_table.add_column("Field name", style="i")
my_table.add_column("Value")
for field in self.__fields__:
attr = getattr(self, field)
if isinstance(attr, str):
attr_str = attr
elif hasattr(attr, "create_renderable"):
attr_str = attr.create_renderable()
elif isinstance(attr, BaseModel):
attr_str = attr.json(option=orjson.orjson.OPT_INDENT_2)
else:
attr_str = str(attr)
my_table.add_row(field, attr_str)
return my_table
def __eq__(self, other):
if self.__class__ != other.__class__:
return False
return self.model_data_hash == other.model_data_hash
def __hash__(self):
return self.model_data_hash
constants: Dict[str, Any]
pydantic-field
¶Value constants for this module.
defaults: Dict[str, Any]
pydantic-field
¶Value defaults for this module.
Config
¶create_renderable(self, **config)
¶Source code in kiara/modules/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
my_table = Table(box=box.MINIMAL, show_header=False)
my_table.add_column("Field name", style="i")
my_table.add_column("Value")
for field in self.__fields__:
attr = getattr(self, field)
if isinstance(attr, str):
attr_str = attr
elif hasattr(attr, "create_renderable"):
attr_str = attr.create_renderable()
elif isinstance(attr, BaseModel):
attr_str = attr.json(option=orjson.orjson.OPT_INDENT_2)
else:
attr_str = str(attr)
my_table.add_row(field, attr_str)
return my_table
get(self, key)
¶Get the value for the specified configuation key.
Source code in kiara/modules/__init__.py
def get(self, key: str) -> Any:
"""Get the value for the specified configuation key."""
if key not in self.__fields__:
raise Exception(
f"No config value '{key}' in module config class '{self.__class__.__name__}'."
)
return getattr(self, key)
requires_config(config=None)
classmethod
¶Return whether this class can be used as-is, or requires configuration before an instance can be created.
Source code in kiara/modules/__init__.py
@classmethod
def requires_config(cls, config: Optional[Mapping[str, Any]] = None) -> bool:
"""Return whether this class can be used as-is, or requires configuration before an instance can be created."""
for field_name, field in cls.__fields__.items():
if field.required and field.default is None:
if config:
if config.get(field_name, None) is None:
return True
else:
return True
return False
Methods¶
create_renderable(self, **config)
¶Source code in kiara/modules/__init__.py
def create_renderable(self, **config) -> RenderableType:
if self._operation is not None:
return self._operation
from kiara.models.module.operation import Operation
self._operation = Operation.create_from_module(self)
return self._operation
get_config_value(self, key)
¶Retrieve the value for a specific configuration option.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
key |
str |
the config key |
required |
Returns:
| Type | Description |
|---|---|
Any |
the value for the provided key |
Source code in kiara/modules/__init__.py
def get_config_value(self, key: str) -> Any:
"""Retrieve the value for a specific configuration option.
Arguments:
key: the config key
Returns:
the value for the provided key
"""
try:
return self.config.get(key)
except Exception:
raise Exception(
f"Error accessing config value '{key}' in module {self.__class__._module_type_name}." # type: ignore
)
is_pipeline()
classmethod
¶Check whether this module type is a pipeline, or not.
Source code in kiara/modules/__init__.py
@classmethod
def is_pipeline(cls) -> bool:
"""Check whether this module type is a pipeline, or not."""
return False
process_step(self, inputs, outputs, job_log)
¶Kick off processing for a specific set of input/outputs.
This method calls the implemented [process][kiara.module.KiaraModule.process] method of the inheriting class, as well as wrapping input/output-data related functionality.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
inputs |
ValueMap |
the input value set |
required |
outputs |
ValueMap |
the output value set |
required |
Source code in kiara/modules/__init__.py
def process_step(
self, inputs: "ValueMap", outputs: "ValueMap", job_log: JobLog
) -> None:
"""Kick off processing for a specific set of input/outputs.
This method calls the implemented [process][kiara.module.KiaraModule.process] method of the inheriting class,
as well as wrapping input/output-data related functionality.
Arguments:
inputs: the input value set
outputs: the output value set
"""
signature = inspect.signature(self.process) # type: ignore
if "job_log" not in signature.parameters.keys():
try:
self.process(inputs=inputs, outputs=outputs) # type: ignore
except Exception as e:
if is_debug():
try:
import traceback
traceback.print_exc()
except Exception:
pass
raise e
else:
try:
self.process(inputs=inputs, outputs=outputs, job_log=job_log) # type: ignore
except Exception as e:
if is_debug():
try:
import traceback
traceback.print_exc()
except Exception:
pass
raise e
ModuleCharacteristics (BaseModel)
pydantic-model
¶
Source code in kiara/modules/__init__.py
class ModuleCharacteristics(BaseModel):
is_idempotent: bool = Field(
description="Whether this module is idempotent (aka always produces the same output with the same inputs.",
default=False,
)
is_internal: bool = Field(
description="Hint for frontends whether this module is used predominantly internally, and users won't need to know of its existence.",
default=False,
)
Modules¶
included_core_modules
special
¶
Modules¶
create_from
¶
CreateFromModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/create_from.py
class CreateFromModule(KiaraModule):
_module_type_name: str = None # type: ignore
_config_cls = CreateFromModuleConfig
@classmethod
def retrieve_supported_create_combinations(cls) -> Iterable[Mapping[str, str]]:
result = []
for attr in dir(cls):
if (
len(attr) <= 16
or not attr.startswith("create__")
or "__from__" not in attr
):
continue
tokens = attr.split("__")
if len(tokens) != 4:
continue
source_type = tokens[3]
target_type = tokens[1]
data = {
"source_type": source_type,
"target_type": target_type,
"func": attr,
}
result.append(data)
return result
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
assert source_type not in ["target", "base_name"]
schema = {
source_type: {"type": source_type, "doc": "The value to render."},
}
return schema
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
self.get_config_value("target_type"): {
"type": self.get_config_value("target_type"),
"doc": "The result value.",
}
}
def process(self, inputs: ValueMap, outputs: ValueMap):
source_type = self.get_config_value("source_type")
target_type = self.get_config_value("target_type")
value = inputs.get_value_obj(source_type)
render_config = inputs.get_value_data("render_config")
func_name = f"render__{source_type}__as__{target_type}"
func = getattr(self, func_name)
# TODO: check function signature is valid
result = func(value=value, render_config=render_config)
outputs.set_value("rendered_value", result)
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶Source code in kiara/modules/included_core_modules/create_from.py
class CreateFromModuleConfig(KiaraModuleConfig):
source_type: str = Field(description="The value type of the source value.")
target_type: str = Field(description="The value type of the target.")
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/create_from.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
assert source_type not in ["target", "base_name"]
schema = {
source_type: {"type": source_type, "doc": "The value to render."},
}
return schema
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/create_from.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
self.get_config_value("target_type"): {
"type": self.get_config_value("target_type"),
"doc": "The result value.",
}
}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/create_from.py
def process(self, inputs: ValueMap, outputs: ValueMap):
source_type = self.get_config_value("source_type")
target_type = self.get_config_value("target_type")
value = inputs.get_value_obj(source_type)
render_config = inputs.get_value_data("render_config")
func_name = f"render__{source_type}__as__{target_type}"
func = getattr(self, func_name)
# TODO: check function signature is valid
result = func(value=value, render_config=render_config)
outputs.set_value("rendered_value", result)
retrieve_supported_create_combinations()
classmethod
¶Source code in kiara/modules/included_core_modules/create_from.py
@classmethod
def retrieve_supported_create_combinations(cls) -> Iterable[Mapping[str, str]]:
result = []
for attr in dir(cls):
if (
len(attr) <= 16
or not attr.startswith("create__")
or "__from__" not in attr
):
continue
tokens = attr.split("__")
if len(tokens) != 4:
continue
source_type = tokens[3]
target_type = tokens[1]
data = {
"source_type": source_type,
"target_type": target_type,
"func": attr,
}
result.append(data)
return result
CreateFromModuleConfig (KiaraModuleConfig)
pydantic-model
¶Source code in kiara/modules/included_core_modules/create_from.py
class CreateFromModuleConfig(KiaraModuleConfig):
source_type: str = Field(description="The value type of the source value.")
target_type: str = Field(description="The value type of the target.")
filesystem
¶
ImportFileBundleModule (KiaraModule)
¶Import a folder (file_bundle) from the local filesystem.
Source code in kiara/modules/included_core_modules/filesystem.py
class ImportFileBundleModule(KiaraModule):
"""Import a folder (file_bundle) from the local filesystem."""
_module_type_name = "import.file_bundle"
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {
"path": {"type": "string", "doc": "The local path of the folder to import."}
}
def create_outputs_schema(
self,
) -> ValueSetSchema:
return {
"file_bundle": {"type": "file_bundle", "doc": "The imported file bundle."}
}
def process(self, inputs: ValueMap, outputs: ValueMap):
path = inputs.get_value_data("path")
file_bundle = FileBundle.import_folder(source=path)
outputs.set_value("file_bundle", file_bundle)
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/filesystem.py
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {
"path": {"type": "string", "doc": "The local path of the folder to import."}
}
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/filesystem.py
def create_outputs_schema(
self,
) -> ValueSetSchema:
return {
"file_bundle": {"type": "file_bundle", "doc": "The imported file bundle."}
}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/filesystem.py
def process(self, inputs: ValueMap, outputs: ValueMap):
path = inputs.get_value_data("path")
file_bundle = FileBundle.import_folder(source=path)
outputs.set_value("file_bundle", file_bundle)
ImportFileModule (KiaraModule)
¶Import a file from the local filesystem.
Source code in kiara/modules/included_core_modules/filesystem.py
class ImportFileModule(KiaraModule):
"""Import a file from the local filesystem."""
_module_type_name = "import.file"
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {"path": {"type": "string", "doc": "The local path to the file."}}
def create_outputs_schema(
self,
) -> ValueSetSchema:
return {"file": {"type": "file", "doc": "The loaded files."}}
def process(self, inputs: ValueMap, outputs: ValueMap):
path = inputs.get_value_data("path")
file = FileModel.load_file(source=path)
outputs.set_value("file", file)
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/filesystem.py
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {"path": {"type": "string", "doc": "The local path to the file."}}
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/filesystem.py
def create_outputs_schema(
self,
) -> ValueSetSchema:
return {"file": {"type": "file", "doc": "The loaded files."}}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/filesystem.py
def process(self, inputs: ValueMap, outputs: ValueMap):
path = inputs.get_value_data("path")
file = FileModel.load_file(source=path)
outputs.set_value("file", file)
LoadFileBundleFromStoreModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/filesystem.py
class LoadFileBundleFromStoreModule(KiaraModule):
_module_type_name = "load.file_bundle.from_data_store"
def _retrieve_module_characteristics(self) -> ModuleCharacteristics:
return ModuleCharacteristics(is_internal=True)
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {
"inline_data": {"type": "any", "doc": "The file bundle metadata."},
"path": {
"type": "string",
"doc": "The path to the provisioned file bundle.",
},
}
def create_outputs_schema(
self,
) -> ValueSetSchema:
return {
"file_bundle": {
"type": "file_bundle",
"doc": "The loaded file_bundle value.",
}
}
def process(self, inputs: ValueMap, outputs: ValueMap):
path = inputs.get_value_data("path")
# bundle_name = inputs.get_value_data("bundle_name")
# import_time_str = inputs.get_value_data("import_time")
# import_time = parser.parse(import_time_str)
bundle_data = inputs.get_value_data("inline_data")
file_bundle = FileBundle(**bundle_data)
file_bundle._path = path
for rel_path, model in file_bundle.included_files.items():
model._path = os.path.join(path, rel_path)
outputs.set_value("file_bundle", file_bundle)
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/filesystem.py
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {
"inline_data": {"type": "any", "doc": "The file bundle metadata."},
"path": {
"type": "string",
"doc": "The path to the provisioned file bundle.",
},
}
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/filesystem.py
def create_outputs_schema(
self,
) -> ValueSetSchema:
return {
"file_bundle": {
"type": "file_bundle",
"doc": "The loaded file_bundle value.",
}
}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/filesystem.py
def process(self, inputs: ValueMap, outputs: ValueMap):
path = inputs.get_value_data("path")
# bundle_name = inputs.get_value_data("bundle_name")
# import_time_str = inputs.get_value_data("import_time")
# import_time = parser.parse(import_time_str)
bundle_data = inputs.get_value_data("inline_data")
file_bundle = FileBundle(**bundle_data)
file_bundle._path = path
for rel_path, model in file_bundle.included_files.items():
model._path = os.path.join(path, rel_path)
outputs.set_value("file_bundle", file_bundle)
LoadFileFromStoreModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/filesystem.py
class LoadFileFromStoreModule(KiaraModule):
_module_type_name = "load.file.from_data_store"
def _retrieve_module_characteristics(self) -> ModuleCharacteristics:
return ModuleCharacteristics(is_internal=True)
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {
"file_name": {"type": "string", "doc": "The name of the file."},
"import_time": {
"type": "string",
"doc": "The (original) import time of the file.",
},
"bytes_structure": {
"type": "any",
"doc": "The bytes that make up the file.",
},
}
def create_outputs_schema(
self,
) -> ValueSetSchema:
return {"file": {"type": "file", "doc": "The loaded files."}}
def process(self, inputs: ValueMap, outputs: ValueMap):
file_name = inputs.get_value_data("file_name")
import_time_str = inputs.get_value_data("import_time")
bytes_structure: BytesStructure = inputs.get_value_data("bytes_structure")
assert len(bytes_structure.chunk_map) == 1
import_time = parser.parse(import_time_str)
file_chunks = bytes_structure.chunk_map[file_name]
assert len(file_chunks) == 1
chunk = file_chunks[0]
assert isinstance(chunk, str)
file = FileModel.load_file(
source=chunk,
file_name=file_name,
import_time=import_time,
)
outputs.set_value("file", file)
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/filesystem.py
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {
"file_name": {"type": "string", "doc": "The name of the file."},
"import_time": {
"type": "string",
"doc": "The (original) import time of the file.",
},
"bytes_structure": {
"type": "any",
"doc": "The bytes that make up the file.",
},
}
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/filesystem.py
def create_outputs_schema(
self,
) -> ValueSetSchema:
return {"file": {"type": "file", "doc": "The loaded files."}}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/filesystem.py
def process(self, inputs: ValueMap, outputs: ValueMap):
file_name = inputs.get_value_data("file_name")
import_time_str = inputs.get_value_data("import_time")
bytes_structure: BytesStructure = inputs.get_value_data("bytes_structure")
assert len(bytes_structure.chunk_map) == 1
import_time = parser.parse(import_time_str)
file_chunks = bytes_structure.chunk_map[file_name]
assert len(file_chunks) == 1
chunk = file_chunks[0]
assert isinstance(chunk, str)
file = FileModel.load_file(
source=chunk,
file_name=file_name,
import_time=import_time,
)
outputs.set_value("file", file)
SaveFileBundleToStoreModule (PersistValueModule)
¶Source code in kiara/modules/included_core_modules/filesystem.py
class SaveFileBundleToStoreModule(PersistValueModule):
_module_type_name = "file_bundle.save_to_data_store"
def get_persistence_target_name(self) -> str:
return "data_store"
def get_persistence_format_name(self) -> str:
return "file_bundle"
def data_type__file_bundle(
self, value: Value, persistence_config: Mapping[str, Any]
) -> Tuple[LoadConfig, BytesStructure]:
"""Persist single files into a local kiara data store."""
file_bundle: FileBundle = value.data
bytes_structure_data: Dict[str, List[Union[str, bytes]]] = {}
for rel_path, file_model in file_bundle.included_files.items():
bytes_structure_data[rel_path] = [file_model.path]
bytes_structure = BytesStructure.construct(
data_type="file_bundle", data_type_config={}, chunk_map=bytes_structure_data
)
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.LINK_FOLDER,
"module_type": "load.file_bundle.from_data_store",
"inputs": {
"inline_data": LOAD_CONFIG_PLACEHOLDER,
"path": LOAD_CONFIG_PLACEHOLDER,
"bytes_structure": LOAD_CONFIG_PLACEHOLDER,
},
"inline_data": file_bundle.dict(),
"output_name": "file_bundle",
}
load_config = LoadConfig(**load_config_data)
return load_config, bytes_structure
data_type__file_bundle(self, value, persistence_config)
¶Persist single files into a local kiara data store.
Source code in kiara/modules/included_core_modules/filesystem.py
def data_type__file_bundle(
self, value: Value, persistence_config: Mapping[str, Any]
) -> Tuple[LoadConfig, BytesStructure]:
"""Persist single files into a local kiara data store."""
file_bundle: FileBundle = value.data
bytes_structure_data: Dict[str, List[Union[str, bytes]]] = {}
for rel_path, file_model in file_bundle.included_files.items():
bytes_structure_data[rel_path] = [file_model.path]
bytes_structure = BytesStructure.construct(
data_type="file_bundle", data_type_config={}, chunk_map=bytes_structure_data
)
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.LINK_FOLDER,
"module_type": "load.file_bundle.from_data_store",
"inputs": {
"inline_data": LOAD_CONFIG_PLACEHOLDER,
"path": LOAD_CONFIG_PLACEHOLDER,
"bytes_structure": LOAD_CONFIG_PLACEHOLDER,
},
"inline_data": file_bundle.dict(),
"output_name": "file_bundle",
}
load_config = LoadConfig(**load_config_data)
return load_config, bytes_structure
get_persistence_format_name(self)
¶Source code in kiara/modules/included_core_modules/filesystem.py
def get_persistence_format_name(self) -> str:
return "file_bundle"
get_persistence_target_name(self)
¶Source code in kiara/modules/included_core_modules/filesystem.py
def get_persistence_target_name(self) -> str:
return "data_store"
SaveFileToStoreModule (PersistValueModule)
¶Source code in kiara/modules/included_core_modules/filesystem.py
class SaveFileToStoreModule(PersistValueModule):
_module_type_name = "file.save_to_data_store"
def get_persistence_target_name(self) -> str:
return "data_store"
def get_persistence_format_name(self) -> str:
return "file"
def data_type__file(
self, value: Value, persistence_config: Mapping[str, Any]
) -> Tuple[LoadConfig, BytesStructure]:
"""Persist single files into a local kiara data store."""
file: FileModel = value.data
bytes_structure_data: Mapping[str, List[Union[str, bytes]]] = {
file.file_name: [file.path]
}
bytes_structure = BytesStructure.construct(
data_type="file", data_type_config={}, chunk_map=bytes_structure_data
)
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.FILE_PATH_MAP,
"module_type": "load.file.from_data_store",
"inputs": {
"file_name": file.file_name,
"import_time": str(file.import_time),
"bytes_structure": LOAD_CONFIG_PLACEHOLDER,
},
"output_name": "file",
}
load_config = LoadConfig(**load_config_data)
return load_config, bytes_structure
data_type__file(self, value, persistence_config)
¶Persist single files into a local kiara data store.
Source code in kiara/modules/included_core_modules/filesystem.py
def data_type__file(
self, value: Value, persistence_config: Mapping[str, Any]
) -> Tuple[LoadConfig, BytesStructure]:
"""Persist single files into a local kiara data store."""
file: FileModel = value.data
bytes_structure_data: Mapping[str, List[Union[str, bytes]]] = {
file.file_name: [file.path]
}
bytes_structure = BytesStructure.construct(
data_type="file", data_type_config={}, chunk_map=bytes_structure_data
)
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.FILE_PATH_MAP,
"module_type": "load.file.from_data_store",
"inputs": {
"file_name": file.file_name,
"import_time": str(file.import_time),
"bytes_structure": LOAD_CONFIG_PLACEHOLDER,
},
"output_name": "file",
}
load_config = LoadConfig(**load_config_data)
return load_config, bytes_structure
get_persistence_format_name(self)
¶Source code in kiara/modules/included_core_modules/filesystem.py
def get_persistence_format_name(self) -> str:
return "file"
get_persistence_target_name(self)
¶Source code in kiara/modules/included_core_modules/filesystem.py
def get_persistence_target_name(self) -> str:
return "data_store"
metadata
¶
ExtractMetadataModule (KiaraModule)
¶Base class to use when writing a module to extract metadata from a file.
It's possible to use any arbitrary kiara module for this purpose, but sub-classing this makes it easier.
Source code in kiara/modules/included_core_modules/metadata.py
class ExtractMetadataModule(KiaraModule):
"""Base class to use when writing a module to extract metadata from a file.
It's possible to use any arbitrary *kiara* module for this purpose, but sub-classing this makes it easier.
"""
_config_cls = MetadataModuleConfig
_module_type_name: str = "value.extract_metadata"
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
data_type_name = self.get_config_value("data_type")
inputs = {
"value": {
"type": data_type_name,
"doc": f"A value of type '{data_type_name}'",
"optional": False,
}
}
return inputs
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
outputs = {
"value_metadata": {
"type": "internal_model",
"doc": "The metadata for the provided value.",
}
}
return outputs
def process(self, inputs: ValueMap, outputs: ValueMap) -> None:
# metadata_key = self.get_config_value("metadata_key")
# data_type = self.get_config_value("data_type")
metadata_model: PythonClass = self.get_config_value("metadata_model")
value = inputs.get_value_obj("value")
metadata_model_cls: Type[ValueMetadata] = metadata_model.get_class() # type: ignore
metadata = metadata_model_cls.create_value_metadata(value=value)
if not isinstance(metadata, metadata_model_cls):
raise KiaraProcessingException(
f"Invalid metadata model result, should be class '{metadata_model_cls.__name__}', but is: {metadata.__class__.__name__}. This is most likely a bug."
)
# if isinstance(metadata, Mapping):
# md = metadata_model_cls(**metadata)
# elif isinstance(metadata, metadata_model_cls):
# md = metadata
# else:
# raise KiaraProcessingException(
# f"Invalid type '{type(metadata)}' for result metadata, must be a mapping or subclass of '{metadata_model_cls.__name__}'."
# )
outputs.set_value("value_metadata", metadata)
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶Source code in kiara/modules/included_core_modules/metadata.py
class MetadataModuleConfig(KiaraModuleConfig):
# metadata_key: str = Field(description="The key for the metadata association.")
data_type: str = Field(description="The data type this module will be used for.")
metadata_model: PythonClass = Field(description="The metadata model class.")
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/metadata.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
data_type_name = self.get_config_value("data_type")
inputs = {
"value": {
"type": data_type_name,
"doc": f"A value of type '{data_type_name}'",
"optional": False,
}
}
return inputs
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/metadata.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
outputs = {
"value_metadata": {
"type": "internal_model",
"doc": "The metadata for the provided value.",
}
}
return outputs
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/metadata.py
def process(self, inputs: ValueMap, outputs: ValueMap) -> None:
# metadata_key = self.get_config_value("metadata_key")
# data_type = self.get_config_value("data_type")
metadata_model: PythonClass = self.get_config_value("metadata_model")
value = inputs.get_value_obj("value")
metadata_model_cls: Type[ValueMetadata] = metadata_model.get_class() # type: ignore
metadata = metadata_model_cls.create_value_metadata(value=value)
if not isinstance(metadata, metadata_model_cls):
raise KiaraProcessingException(
f"Invalid metadata model result, should be class '{metadata_model_cls.__name__}', but is: {metadata.__class__.__name__}. This is most likely a bug."
)
# if isinstance(metadata, Mapping):
# md = metadata_model_cls(**metadata)
# elif isinstance(metadata, metadata_model_cls):
# md = metadata
# else:
# raise KiaraProcessingException(
# f"Invalid type '{type(metadata)}' for result metadata, must be a mapping or subclass of '{metadata_model_cls.__name__}'."
# )
outputs.set_value("value_metadata", metadata)
MetadataModuleConfig (KiaraModuleConfig)
pydantic-model
¶Source code in kiara/modules/included_core_modules/metadata.py
class MetadataModuleConfig(KiaraModuleConfig):
# metadata_key: str = Field(description="The key for the metadata association.")
data_type: str = Field(description="The data type this module will be used for.")
metadata_model: PythonClass = Field(description="The metadata model class.")
persistence
¶
LoadDataModuleConfig (KiaraModuleConfig)
pydantic-model
¶
LoadInlineDataModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/persistence.py
class LoadInlineDataModule(KiaraModule):
_module_type_name = "value.load_inline"
_config_cls = LoadDataModuleConfig
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {
"inline_data": {
"type": "any",
"doc": "The data.",
}
}
def create_outputs_schema(
self,
) -> ValueSetSchema:
data_type = self.get_config_value("data_type")
return {
"value": {"type": data_type, "doc": f"The deserialized {data_type} value."}
}
def process(self, inputs: ValueMap, outputs: ValueMap):
data = inputs.get_value_data("inline_data")
outputs.set_value("value", data)
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/persistence.py
def create_inputs_schema(
self,
) -> ValueSetSchema:
return {
"inline_data": {
"type": "any",
"doc": "The data.",
}
}
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/persistence.py
def create_outputs_schema(
self,
) -> ValueSetSchema:
data_type = self.get_config_value("data_type")
return {
"value": {"type": data_type, "doc": f"The deserialized {data_type} value."}
}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/persistence.py
def process(self, inputs: ValueMap, outputs: ValueMap):
data = inputs.get_value_data("inline_data")
outputs.set_value("value", data)
LoadInternalModelModule (KiaraModule)
¶Load a json file from disk and create a kiara value from it.
Source code in kiara/modules/included_core_modules/persistence.py
class LoadInternalModelModule(KiaraModule):
"""Load a json file from disk and create a kiara value from it."""
_module_type_name = "internal_model.load_from_store"
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"model_data": {
"type": "any",
"doc": "The serialized model data.",
}
}
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"internal_model": {
"type": "internal_model",
"doc": "The deserialized internal_model value, loaded from the data store.",
}
}
def _retrieve_module_characteristics(self) -> ModuleCharacteristics:
return ModuleCharacteristics(is_internal=True)
def process(self, inputs: ValueMap, outputs: ValueMap):
data_str = inputs.get_value_data("model_data")
data = orjson.loads(data_str)
model_data = data["data"]
python_class_data = data["python_class"]
model_cls = PythonClass(**python_class_data).get_class()
model = model_cls(**model_data)
outputs.set_value("internal_model", model)
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/persistence.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"model_data": {
"type": "any",
"doc": "The serialized model data.",
}
}
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/persistence.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"internal_model": {
"type": "internal_model",
"doc": "The deserialized internal_model value, loaded from the data store.",
}
}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/persistence.py
def process(self, inputs: ValueMap, outputs: ValueMap):
data_str = inputs.get_value_data("model_data")
data = orjson.loads(data_str)
model_data = data["data"]
python_class_data = data["python_class"]
model_cls = PythonClass(**python_class_data).get_class()
model = model_cls(**model_data)
outputs.set_value("internal_model", model)
LoadPickledDataModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/persistence.py
class LoadPickledDataModule(KiaraModule):
_module_type_name = "value.load_pickled_data"
_config_cls = LoadDataModuleConfig
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {"bytes_structure": {"type": "any", "doc": "The raw pickle data."}}
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
data_type_name = self.get_config_value("data_type")
return {
data_type_name: {
"type": data_type_name,
"doc": f"The deserialized {data_type_name} value, loaded from disk.",
}
}
def process(self, inputs: ValueMap, outputs: ValueMap):
try:
import pickle5 as pickle
except Exception:
import pickle # type: ignore
data_type_name = self.get_config_value("data_type")
bytes_structure: BytesStructure = inputs.get_value_data("bytes_structure")
assert len(bytes_structure.chunk_map) == 1
key = next(iter(bytes_structure.chunk_map))
value = bytes_structure.chunk_map[key]
assert len(bytes_structure.chunk_map) == 1
data = pickle.loads(value[0])
outputs.set_value(data_type_name, data)
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/persistence.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {"bytes_structure": {"type": "any", "doc": "The raw pickle data."}}
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/persistence.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
data_type_name = self.get_config_value("data_type")
return {
data_type_name: {
"type": data_type_name,
"doc": f"The deserialized {data_type_name} value, loaded from disk.",
}
}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/persistence.py
def process(self, inputs: ValueMap, outputs: ValueMap):
try:
import pickle5 as pickle
except Exception:
import pickle # type: ignore
data_type_name = self.get_config_value("data_type")
bytes_structure: BytesStructure = inputs.get_value_data("bytes_structure")
assert len(bytes_structure.chunk_map) == 1
key = next(iter(bytes_structure.chunk_map))
value = bytes_structure.chunk_map[key]
assert len(bytes_structure.chunk_map) == 1
data = pickle.loads(value[0])
outputs.set_value(data_type_name, data)
PersistValueConfig (KiaraModuleConfig)
pydantic-model
¶Source code in kiara/modules/included_core_modules/persistence.py
class PersistValueConfig(KiaraModuleConfig):
source_type: str = Field(description="The value type of the source.")
source_type_config: Dict[str, Any] = Field(
description="The value type config (if applicable).", default_factory=dict
)
@validator("source_type")
def validate_source_type(cls, value):
if value == "persist_config":
raise ValueError(f"Invalid source type: {value}.")
return value
source_type: str
pydantic-field
required
¶The value type of the source.
source_type_config: Dict[str, Any]
pydantic-field
¶The value type config (if applicable).
validate_source_type(value)
classmethod
¶Source code in kiara/modules/included_core_modules/persistence.py
@validator("source_type")
def validate_source_type(cls, value):
if value == "persist_config":
raise ValueError(f"Invalid source type: {value}.")
return value
PersistValueModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/persistence.py
class PersistValueModule(KiaraModule):
_config_cls = PersistValueConfig
@classmethod
def retrieve_supported_source_types(cls) -> Iterable[str]:
result = []
for attr in dir(cls):
if attr.startswith("data_type__"):
result.append(attr[11:])
return result
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
assert source_type not in ["target", "base_name"]
schema = {
source_type: {"type": source_type, "doc": "The value to serialize."},
}
return schema
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"load_config": {
"type": LOAD_CONFIG_DATA_TYPE_NAME,
"type_config": {
"persistence_target": self.get_persistence_target_name(),
"persistence_format": self.get_persistence_format_name(),
},
"doc": "The value in serialized form.",
},
"bytes_structure": {
"type": "any",
"doc": "The actual serialized value bytes or path to saved files containing the value data.",
"optional": True,
},
}
@abc.abstractmethod
def get_persistence_target_name(self) -> str:
pass
@abc.abstractmethod
def get_persistence_format_name(self) -> str:
pass
def process(self, inputs: ValueMap, outputs: ValueMap) -> None:
source_type = self.get_config_value("source_type")
value = inputs.get_value_obj(source_type)
func_name = f"data_type__{self.get_config_value('source_type')}"
func = getattr(self, func_name)
result: LoadConfig
bytes_structure: Optional[BytesStructure]
result, bytes_structure = func(value=value, persistence_config={"x": "y"})
outputs.set_values(load_config=result, bytes_structure=bytes_structure)
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶Source code in kiara/modules/included_core_modules/persistence.py
class PersistValueConfig(KiaraModuleConfig):
source_type: str = Field(description="The value type of the source.")
source_type_config: Dict[str, Any] = Field(
description="The value type config (if applicable).", default_factory=dict
)
@validator("source_type")
def validate_source_type(cls, value):
if value == "persist_config":
raise ValueError(f"Invalid source type: {value}.")
return value
source_type: str
pydantic-field
required
¶The value type of the source.
source_type_config: Dict[str, Any]
pydantic-field
¶The value type config (if applicable).
validate_source_type(value)
classmethod
¶Source code in kiara/modules/included_core_modules/persistence.py
@validator("source_type")
def validate_source_type(cls, value):
if value == "persist_config":
raise ValueError(f"Invalid source type: {value}.")
return value
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/persistence.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
assert source_type not in ["target", "base_name"]
schema = {
source_type: {"type": source_type, "doc": "The value to serialize."},
}
return schema
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/persistence.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"load_config": {
"type": LOAD_CONFIG_DATA_TYPE_NAME,
"type_config": {
"persistence_target": self.get_persistence_target_name(),
"persistence_format": self.get_persistence_format_name(),
},
"doc": "The value in serialized form.",
},
"bytes_structure": {
"type": "any",
"doc": "The actual serialized value bytes or path to saved files containing the value data.",
"optional": True,
},
}
get_persistence_format_name(self)
¶Source code in kiara/modules/included_core_modules/persistence.py
@abc.abstractmethod
def get_persistence_format_name(self) -> str:
pass
get_persistence_target_name(self)
¶Source code in kiara/modules/included_core_modules/persistence.py
@abc.abstractmethod
def get_persistence_target_name(self) -> str:
pass
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/persistence.py
def process(self, inputs: ValueMap, outputs: ValueMap) -> None:
source_type = self.get_config_value("source_type")
value = inputs.get_value_obj(source_type)
func_name = f"data_type__{self.get_config_value('source_type')}"
func = getattr(self, func_name)
result: LoadConfig
bytes_structure: Optional[BytesStructure]
result, bytes_structure = func(value=value, persistence_config={"x": "y"})
outputs.set_values(load_config=result, bytes_structure=bytes_structure)
retrieve_supported_source_types()
classmethod
¶Source code in kiara/modules/included_core_modules/persistence.py
@classmethod
def retrieve_supported_source_types(cls) -> Iterable[str]:
result = []
for attr in dir(cls):
if attr.startswith("data_type__"):
result.append(attr[11:])
return result
SaveInlineDataModule (PersistValueModule)
¶Source code in kiara/modules/included_core_modules/persistence.py
class SaveInlineDataModule(PersistValueModule):
_module_type_name: str = None # type: ignore
def get_persistence_target_name(self) -> str:
return "inline"
def get_persistence_format_name(self) -> str:
return "json_string"
def create_inline_load_config(self, data: Any, data_type: str):
try:
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.INLINE,
"module_type": "value.load_inline",
"module_config": {"data_type": data_type},
"inputs": {"inline_data": LOAD_CONFIG_PLACEHOLDER},
"inline_data": data,
"output_name": "value",
}
load_config = LoadConfig(**load_config_data)
except Exception as e:
raise KiaraProcessingException(
f"Can't serialize value of type '{data_type}' to json: {e}."
)
return load_config
create_inline_load_config(self, data, data_type)
¶Source code in kiara/modules/included_core_modules/persistence.py
def create_inline_load_config(self, data: Any, data_type: str):
try:
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.INLINE,
"module_type": "value.load_inline",
"module_config": {"data_type": data_type},
"inputs": {"inline_data": LOAD_CONFIG_PLACEHOLDER},
"inline_data": data,
"output_name": "value",
}
load_config = LoadConfig(**load_config_data)
except Exception as e:
raise KiaraProcessingException(
f"Can't serialize value of type '{data_type}' to json: {e}."
)
return load_config
get_persistence_format_name(self)
¶Source code in kiara/modules/included_core_modules/persistence.py
def get_persistence_format_name(self) -> str:
return "json_string"
get_persistence_target_name(self)
¶Source code in kiara/modules/included_core_modules/persistence.py
def get_persistence_target_name(self) -> str:
return "inline"
SaveInternalModelModule (PersistValueModule)
¶Persist internally used model data.
Source code in kiara/modules/included_core_modules/persistence.py
class SaveInternalModelModule(PersistValueModule):
"""Persist internally used model data."""
_module_type_name = "internal_model.save_to.disk.as.json_file"
def get_persistence_target_name(self) -> str:
return "disk"
def get_persistence_format_name(self) -> str:
return "json_file"
def data_type__internal_model(
self, value: Value, persistence_config: Mapping[str, Any]
) -> Tuple[LoadConfig, Optional[BytesStructure]]:
"""Persist internally used model data as a json file."""
try:
value_metadata: KiaraModel = value.data
data_json = value_metadata.json()
python_class_json = PythonClass.from_class(value_metadata.__class__).json()
all_json = (
'{"data": ' + data_json + ', "python_class": ' + python_class_json + "}"
)
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.INLINE,
"module_type": "internal_model.load_from_store",
"inputs": {"model_data": all_json},
"output_name": "internal_model",
}
load_config = LoadConfig(**load_config_data)
except Exception as e:
raise KiaraProcessingException(
f"Can't serialize value of type '{value.value_schema.type}' to json: {e}."
)
return load_config, None
data_type__internal_model(self, value, persistence_config)
¶Persist internally used model data as a json file.
Source code in kiara/modules/included_core_modules/persistence.py
def data_type__internal_model(
self, value: Value, persistence_config: Mapping[str, Any]
) -> Tuple[LoadConfig, Optional[BytesStructure]]:
"""Persist internally used model data as a json file."""
try:
value_metadata: KiaraModel = value.data
data_json = value_metadata.json()
python_class_json = PythonClass.from_class(value_metadata.__class__).json()
all_json = (
'{"data": ' + data_json + ', "python_class": ' + python_class_json + "}"
)
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.INLINE,
"module_type": "internal_model.load_from_store",
"inputs": {"model_data": all_json},
"output_name": "internal_model",
}
load_config = LoadConfig(**load_config_data)
except Exception as e:
raise KiaraProcessingException(
f"Can't serialize value of type '{value.value_schema.type}' to json: {e}."
)
return load_config, None
get_persistence_format_name(self)
¶Source code in kiara/modules/included_core_modules/persistence.py
def get_persistence_format_name(self) -> str:
return "json_file"
get_persistence_target_name(self)
¶Source code in kiara/modules/included_core_modules/persistence.py
def get_persistence_target_name(self) -> str:
return "disk"
SavePickleToDiskModule (PersistValueModule)
¶Source code in kiara/modules/included_core_modules/persistence.py
class SavePickleToDiskModule(PersistValueModule):
_module_type_name = "value.save_to.disk.as.pickle"
def get_persistence_target_name(self) -> str:
return "disk"
def get_persistence_format_name(self) -> str:
return "pickle_file"
def data_type__any(
self, value: Value, persistence_config: Mapping[str, Any]
) -> Tuple[LoadConfig, Optional[BytesStructure]]:
"""Persist any Python object using 'pickle'."""
try:
import pickle5 as pickle
except Exception:
import pickle # type: ignore
pickled_bytes = pickle.dumps(value.data, protocol=5)
bytes_structure_data: Dict[str, Any] = {
"data_type": value.value_schema.type,
"data_type_config": value.value_schema.type_config,
"chunk_map": {"serialized_value.pickle": [pickled_bytes]},
}
bytes_structure = BytesStructure.construct(**bytes_structure_data)
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.BYTES,
"module_type": "value.load_pickled_data",
"module_config": {
"data_type": value.value_schema.type,
},
"inputs": {"bytes_structure": LOAD_CONFIG_PLACEHOLDER},
"output_name": value.value_schema.type,
}
load_config = LoadConfig(**load_config_data)
return load_config, bytes_structure
data_type__any(self, value, persistence_config)
¶Persist any Python object using 'pickle'.
Source code in kiara/modules/included_core_modules/persistence.py
def data_type__any(
self, value: Value, persistence_config: Mapping[str, Any]
) -> Tuple[LoadConfig, Optional[BytesStructure]]:
"""Persist any Python object using 'pickle'."""
try:
import pickle5 as pickle
except Exception:
import pickle # type: ignore
pickled_bytes = pickle.dumps(value.data, protocol=5)
bytes_structure_data: Dict[str, Any] = {
"data_type": value.value_schema.type,
"data_type_config": value.value_schema.type_config,
"chunk_map": {"serialized_value.pickle": [pickled_bytes]},
}
bytes_structure = BytesStructure.construct(**bytes_structure_data)
load_config_data = {
"provisioning_strategy": ByteProvisioningStrategy.BYTES,
"module_type": "value.load_pickled_data",
"module_config": {
"data_type": value.value_schema.type,
},
"inputs": {"bytes_structure": LOAD_CONFIG_PLACEHOLDER},
"output_name": value.value_schema.type,
}
load_config = LoadConfig(**load_config_data)
return load_config, bytes_structure
get_persistence_format_name(self)
¶Source code in kiara/modules/included_core_modules/persistence.py
def get_persistence_format_name(self) -> str:
return "pickle_file"
get_persistence_target_name(self)
¶Source code in kiara/modules/included_core_modules/persistence.py
def get_persistence_target_name(self) -> str:
return "disk"
SaveStringModule (SaveInlineDataModule)
¶Source code in kiara/modules/included_core_modules/persistence.py
class SaveStringModule(SaveInlineDataModule):
_module_type_name = "string.save_inline"
def data_type__string(self, value: Value, persistence_config: Mapping[str, Any]):
load_config = self.create_inline_load_config(
data=value.data, data_type=value.value_schema.type
)
return load_config, None
data_type__string(self, value, persistence_config)
¶Source code in kiara/modules/included_core_modules/persistence.py
def data_type__string(self, value: Value, persistence_config: Mapping[str, Any]):
load_config = self.create_inline_load_config(
data=value.data, data_type=value.value_schema.type
)
return load_config, None
pipeline
¶
PipelineModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/pipeline.py
class PipelineModule(KiaraModule):
_config_cls = PipelineConfig
_module_type_name = "pipeline"
def __init__(
self,
module_config: Union[None, KIARA_CONFIG, Mapping[str, Any]] = None,
):
self._job_registry: Optional[JobRegistry] = None
super().__init__(module_config=module_config)
@classmethod
def is_pipeline(cls) -> bool:
return True
def _set_job_registry(self, job_registry: "JobRegistry"):
self._job_registry = job_registry
def create_inputs_schema(
self,
) -> ValueSetSchema:
pipeline_structure: PipelineStructure = self.config.structure
return pipeline_structure.pipeline_inputs_schema
def create_outputs_schema(
self,
) -> ValueSetSchema:
pipeline_structure: PipelineStructure = self.config.structure
return pipeline_structure.pipeline_outputs_schema
def process(self, inputs: ValueMap, outputs: ValueMapWritable, job_log: JobLog):
pipeline_structure: PipelineStructure = self.config.structure
pipeline = Pipeline(
structure=pipeline_structure, data_registry=outputs._data_registry
)
assert self._job_registry is not None
controller = SinglePipelineBatchController(
pipeline=pipeline, job_registry=self._job_registry
)
pipeline.set_pipeline_inputs(inputs=inputs)
controller.process_pipeline()
# TODO: resolve values first?
outputs.set_values(**pipeline.get_current_pipeline_outputs())
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶A class to hold the configuration for a [PipelineModule][kiara.pipeline.module.PipelineModule].
If you want to control the pipeline input and output names, you need to have to provide a map that uses the autogenerated field name ([step_id]__[alias] -- 2 underscores!!) as key, and the desired field name as value. The reason that schema for the autogenerated field names exist is that it's hard to ensure the uniqueness of each field; some steps can have the same input field names, but will need different input values. In some cases, some inputs of different steps need the same input. Those sorts of things. So, to make sure that we always use the right values, I chose to implement a conservative default approach, accepting that in some cases the user will be prompted for duplicate inputs for the same value.
To remedy that, the pipeline creator has the option to manually specify a mapping to rename some or all of the input/output fields.
Further, because in a lot of cases there won't be any overlapping fields, the creator can specify auto,
in which case Kiara will automatically create a mapping that tries to map autogenerated field names
to the shortest possible names for each case.
Examples:
Configuration for a pipeline module that functions as a nand logic gate (in Python):
and_step = PipelineStepConfig(module_type="and", step_id="and")
not_step = PipelineStepConfig(module_type="not", step_id="not", input_links={"a": ["and.y"]}
nand_p_conf = PipelineConfig(doc="Returns 'False' if both inputs are 'True'.",
steps=[and_step, not_step],
input_aliases={
"and__a": "a",
"and__b": "b"
},
output_aliases={
"not__y": "y"
}}
Or, the same thing in json:
{
"module_type_name": "nand",
"doc": "Returns 'False' if both inputs are 'True'.",
"steps": [
{
"module_type": "and",
"step_id": "and"
},
{
"module_type": "not",
"step_id": "not",
"input_links": {
"a": "and.y"
}
}
],
"input_aliases": {
"and__a": "a",
"and__b": "b"
},
"output_aliases": {
"not__y": "y"
}
}
Source code in kiara/modules/included_core_modules/pipeline.py
class PipelineConfig(KiaraModuleConfig):
"""A class to hold the configuration for a [PipelineModule][kiara.pipeline.module.PipelineModule].
If you want to control the pipeline input and output names, you need to have to provide a map that uses the
autogenerated field name ([step_id]__[alias] -- 2 underscores!!) as key, and the desired field name
as value. The reason that schema for the autogenerated field names exist is that it's hard to ensure
the uniqueness of each field; some steps can have the same input field names, but will need different input
values. In some cases, some inputs of different steps need the same input. Those sorts of things.
So, to make sure that we always use the right values, I chose to implement a conservative default approach,
accepting that in some cases the user will be prompted for duplicate inputs for the same value.
To remedy that, the pipeline creator has the option to manually specify a mapping to rename some or all of
the input/output fields.
Further, because in a lot of cases there won't be any overlapping fields, the creator can specify ``auto``,
in which case *Kiara* will automatically create a mapping that tries to map autogenerated field names
to the shortest possible names for each case.
Examples:
Configuration for a pipeline module that functions as a ``nand`` logic gate (in Python):
``` python
and_step = PipelineStepConfig(module_type="and", step_id="and")
not_step = PipelineStepConfig(module_type="not", step_id="not", input_links={"a": ["and.y"]}
nand_p_conf = PipelineConfig(doc="Returns 'False' if both inputs are 'True'.",
steps=[and_step, not_step],
input_aliases={
"and__a": "a",
"and__b": "b"
},
output_aliases={
"not__y": "y"
}}
```
Or, the same thing in json:
``` json
{
"module_type_name": "nand",
"doc": "Returns 'False' if both inputs are 'True'.",
"steps": [
{
"module_type": "and",
"step_id": "and"
},
{
"module_type": "not",
"step_id": "not",
"input_links": {
"a": "and.y"
}
}
],
"input_aliases": {
"and__a": "a",
"and__b": "b"
},
"output_aliases": {
"not__y": "y"
}
}
```
"""
@classmethod
def from_file(
cls,
path: str,
kiara: Optional["Kiara"] = None,
# module_map: Optional[Mapping[str, Any]] = None,
):
data = get_data_from_file(path)
pipeline_id = data.pop("pipeline_id", None)
if pipeline_id is None:
pipeline_id = os.path.basename(path)
return cls.from_config(pipeline_id=pipeline_id, data=data, kiara=kiara)
@classmethod
def from_config(
cls,
pipeline_id: str,
data: Mapping[str, Any],
kiara: Optional["Kiara"] = None,
# module_map: Optional[Mapping[str, Any]] = None,
):
if kiara is None:
from kiara.context import Kiara
kiara = Kiara.instance()
if not kiara.operation_registry.is_initialized:
kiara.operation_registry.operations # noqa
return cls._from_config(pipeline_id=pipeline_id, data=data, kiara=kiara)
@classmethod
def _from_config(
cls,
pipeline_id: str,
data: Mapping[str, Any],
kiara: "Kiara",
module_map: Optional[Mapping[str, Any]] = None,
):
data = dict(data)
steps = data.pop("steps")
steps = PipelineStep.create_steps(*steps, kiara=kiara, module_map=module_map)
data["steps"] = steps
if not data.get("input_aliases"):
data["input_aliases"] = create_input_alias_map(steps)
if not data.get("output_aliases"):
data["output_aliases"] = create_output_alias_map(steps)
result = cls(pipeline_id=pipeline_id, **data)
return result
class Config:
extra = Extra.allow
validate_assignment = True
pipeline_id: str = Field(description="The name of this pipeline.")
steps: List[PipelineStep] = Field(
description="A list of steps/modules of this pipeline, and their connections.",
)
input_aliases: Dict[str, str] = Field(
description="A map of input aliases, with the calculated (<step_id>__<input_name> -- double underscore!) name as key, and a string (the resulting workflow input alias) as value. Check the documentation for the config class for which marker strings can be used to automatically create this map if possible.",
)
output_aliases: Dict[str, str] = Field(
description="A map of output aliases, with the calculated (<step_id>__<output_name> -- double underscore!) name as key, and a string (the resulting workflow output alias) as value. Check the documentation for the config class for which marker strings can be used to automatically create this map if possible.",
)
documentation: str = Field(
default="-- n/a --", description="Documentation about what the pipeline does."
)
context: Dict[str, Any] = Field(
default_factory=dict, description="Metadata for this workflow."
)
_structure: Optional["PipelineStructure"] = PrivateAttr(default=None)
def _retrieve_id(self) -> str:
return str(self.model_data_hash)
def _retrieve_category_id(self) -> str:
return PIPELINE_CONFIG_TYPE_CATEGORY_ID
def _retrieve_data_to_hash(self) -> Any:
return self.dict()
@validator("steps", pre=True)
def _validate_steps(cls, v):
if not v:
raise ValueError(f"Invalid type for 'steps' value: {type(v)}")
steps = []
for step in v:
if not step:
raise ValueError("No step data provided.")
if isinstance(step, PipelineStep):
steps.append(step)
elif isinstance(step, Mapping):
steps.append(PipelineStep(**step))
else:
raise TypeError(step)
return steps
@property
def structure(self) -> "PipelineStructure":
if self._structure is not None:
return self._structure
from kiara.models.module.pipeline.structure import PipelineStructure
structure = PipelineStructure(pipeline_config=self)
return structure
def create_renderable(self, **config: Any) -> RenderableType:
return create_table_from_model_object(self, exclude_fields={"steps"})
# def create_input_alias_map(self) -> Dict[str, str]:
#
# aliases: Dict[str, List[str]] = {}
# for step in self.steps:
# field_names = step.module.input_names
# for field_name in field_names:
# aliases.setdefault(field_name, []).append(step.step_id)
#
# result: Dict[str, str] = {}
# for field_name, step_ids in aliases.items():
# for step_id in step_ids:
# generated = generate_pipeline_endpoint_name(step_id, field_name)
# result[generated] = generated
#
# return result
#
# def create_output_alias_map(self) -> Dict[str, str]:
#
# aliases: Dict[str, List[str]] = {}
# for step in self.steps:
# field_names = step.module.input_names
# for field_name in field_names:
# aliases.setdefault(field_name, []).append(step.step_id)
#
# result: Dict[str, str] = {}
# for field_name, step_ids in aliases.items():
# for step_id in step_ids:
# generated = generate_pipeline_endpoint_name(step_id, field_name)
# result[generated] = generated
#
# return result
context: Dict[str, Any]
pydantic-field
¶Metadata for this workflow.
documentation: str
pydantic-field
¶Documentation about what the pipeline does.
input_aliases: Dict[str, str]
pydantic-field
required
¶A map of input aliases, with the calculated (
output_aliases: Dict[str, str]
pydantic-field
required
¶A map of output aliases, with the calculated (
pipeline_id: str
pydantic-field
required
¶The name of this pipeline.
steps: List[kiara.models.module.pipeline.PipelineStep]
pydantic-field
required
¶A list of steps/modules of this pipeline, and their connections.
structure: PipelineStructure
property
readonly
¶
Config
¶create_renderable(self, **config)
¶Source code in kiara/modules/included_core_modules/pipeline.py
def create_renderable(self, **config: Any) -> RenderableType:
return create_table_from_model_object(self, exclude_fields={"steps"})
from_config(pipeline_id, data, kiara=None)
classmethod
¶Source code in kiara/modules/included_core_modules/pipeline.py
@classmethod
def from_config(
cls,
pipeline_id: str,
data: Mapping[str, Any],
kiara: Optional["Kiara"] = None,
# module_map: Optional[Mapping[str, Any]] = None,
):
if kiara is None:
from kiara.context import Kiara
kiara = Kiara.instance()
if not kiara.operation_registry.is_initialized:
kiara.operation_registry.operations # noqa
return cls._from_config(pipeline_id=pipeline_id, data=data, kiara=kiara)
from_file(path, kiara=None)
classmethod
¶Source code in kiara/modules/included_core_modules/pipeline.py
@classmethod
def from_file(
cls,
path: str,
kiara: Optional["Kiara"] = None,
# module_map: Optional[Mapping[str, Any]] = None,
):
data = get_data_from_file(path)
pipeline_id = data.pop("pipeline_id", None)
if pipeline_id is None:
pipeline_id = os.path.basename(path)
return cls.from_config(pipeline_id=pipeline_id, data=data, kiara=kiara)
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/pipeline.py
def create_inputs_schema(
self,
) -> ValueSetSchema:
pipeline_structure: PipelineStructure = self.config.structure
return pipeline_structure.pipeline_inputs_schema
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/pipeline.py
def create_outputs_schema(
self,
) -> ValueSetSchema:
pipeline_structure: PipelineStructure = self.config.structure
return pipeline_structure.pipeline_outputs_schema
is_pipeline()
classmethod
¶Check whether this module type is a pipeline, or not.
Source code in kiara/modules/included_core_modules/pipeline.py
@classmethod
def is_pipeline(cls) -> bool:
return True
process(self, inputs, outputs, job_log)
¶Source code in kiara/modules/included_core_modules/pipeline.py
def process(self, inputs: ValueMap, outputs: ValueMapWritable, job_log: JobLog):
pipeline_structure: PipelineStructure = self.config.structure
pipeline = Pipeline(
structure=pipeline_structure, data_registry=outputs._data_registry
)
assert self._job_registry is not None
controller = SinglePipelineBatchController(
pipeline=pipeline, job_registry=self._job_registry
)
pipeline.set_pipeline_inputs(inputs=inputs)
controller.process_pipeline()
# TODO: resolve values first?
outputs.set_values(**pipeline.get_current_pipeline_outputs())
render_value
¶
RenderAnyValueModule (RenderValueModule)
¶Source code in kiara/modules/included_core_modules/render_value.py
class RenderAnyValueModule(RenderValueModule):
_module_type_name = "value.render.any"
def render__any__as__string(self, value: Value, render_config: Dict[str, Any]):
data = value.data
if isinstance(data, KiaraModel):
return data.json(option=orjson.OPT_INDENT_2)
else:
return str(data)
def render__any__as__terminal_renderable(
self, value: Value, render_config: Dict[str, Any]
):
data = value.data
if isinstance(data, BaseModel):
rendered = create_table_from_model_object(
model=data, render_config=render_config
)
elif isinstance(data, Iterable):
rendered = pprint.pformat(data)
else:
rendered = str(data)
return rendered
render__any__as__string(self, value, render_config)
¶Source code in kiara/modules/included_core_modules/render_value.py
def render__any__as__string(self, value: Value, render_config: Dict[str, Any]):
data = value.data
if isinstance(data, KiaraModel):
return data.json(option=orjson.OPT_INDENT_2)
else:
return str(data)
render__any__as__terminal_renderable(self, value, render_config)
¶Source code in kiara/modules/included_core_modules/render_value.py
def render__any__as__terminal_renderable(
self, value: Value, render_config: Dict[str, Any]
):
data = value.data
if isinstance(data, BaseModel):
rendered = create_table_from_model_object(
model=data, render_config=render_config
)
elif isinstance(data, Iterable):
rendered = pprint.pformat(data)
else:
rendered = str(data)
return rendered
RenderValueConfig (KiaraModuleConfig)
pydantic-model
¶Source code in kiara/modules/included_core_modules/render_value.py
class RenderValueConfig(KiaraModuleConfig):
source_type: str = Field(description="The value type of the source value.")
target_type: str = Field(description="The value type of the rendered value.")
@validator("source_type")
def validate_source_type(cls, value):
if value == "render_config":
raise ValueError(f"Invalid source type: {value}.")
return value
source_type: str
pydantic-field
required
¶The value type of the source value.
target_type: str
pydantic-field
required
¶The value type of the rendered value.
validate_source_type(value)
classmethod
¶Source code in kiara/modules/included_core_modules/render_value.py
@validator("source_type")
def validate_source_type(cls, value):
if value == "render_config":
raise ValueError(f"Invalid source type: {value}.")
return value
RenderValueModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/render_value.py
class RenderValueModule(KiaraModule):
_module_type_name: str = None # type: ignore
_config_cls = RenderValueConfig
@classmethod
def retrieve_supported_render_combinations(cls) -> Iterable[Tuple[str, str]]:
result = []
for attr in dir(cls):
if (
len(attr) <= 16
or not attr.startswith("render__")
or "__as__" not in attr
):
continue
attr = attr[8:]
end_start_type = attr.find("__as__")
source_type = attr[0:end_start_type]
target_type = attr[end_start_type + 6 :] # noqa
result.append((source_type, target_type))
return result
# def create_persistence_config_schema(self) -> Optional[Mapping[str, Mapping[str, Any]]]:
# return None
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
assert source_type not in ["target", "base_name"]
schema = {
source_type: {"type": source_type, "doc": "The value to render."},
"render_config": {
"type": "any",
"doc": "Value type dependent render configuration.",
"optional": True,
},
}
return schema
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"rendered_value": {
"type": self.get_config_value("target_type"),
"doc": "The rendered value.",
}
}
def process(self, inputs: ValueMap, outputs: ValueMap):
source_type = self.get_config_value("source_type")
target_type = self.get_config_value("target_type")
value = inputs.get_value_obj(source_type)
render_config = inputs.get_value_data("render_config")
func_name = f"render__{source_type}__as__{target_type}"
func = getattr(self, func_name)
# TODO: check function signature is valid
result = func(value=value, render_config=render_config)
outputs.set_value("rendered_value", result)
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶Source code in kiara/modules/included_core_modules/render_value.py
class RenderValueConfig(KiaraModuleConfig):
source_type: str = Field(description="The value type of the source value.")
target_type: str = Field(description="The value type of the rendered value.")
@validator("source_type")
def validate_source_type(cls, value):
if value == "render_config":
raise ValueError(f"Invalid source type: {value}.")
return value
source_type: str
pydantic-field
required
¶The value type of the source value.
target_type: str
pydantic-field
required
¶The value type of the rendered value.
validate_source_type(value)
classmethod
¶Source code in kiara/modules/included_core_modules/render_value.py
@validator("source_type")
def validate_source_type(cls, value):
if value == "render_config":
raise ValueError(f"Invalid source type: {value}.")
return value
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/render_value.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
assert source_type not in ["target", "base_name"]
schema = {
source_type: {"type": source_type, "doc": "The value to render."},
"render_config": {
"type": "any",
"doc": "Value type dependent render configuration.",
"optional": True,
},
}
return schema
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/render_value.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"rendered_value": {
"type": self.get_config_value("target_type"),
"doc": "The rendered value.",
}
}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/render_value.py
def process(self, inputs: ValueMap, outputs: ValueMap):
source_type = self.get_config_value("source_type")
target_type = self.get_config_value("target_type")
value = inputs.get_value_obj(source_type)
render_config = inputs.get_value_data("render_config")
func_name = f"render__{source_type}__as__{target_type}"
func = getattr(self, func_name)
# TODO: check function signature is valid
result = func(value=value, render_config=render_config)
outputs.set_value("rendered_value", result)
retrieve_supported_render_combinations()
classmethod
¶Source code in kiara/modules/included_core_modules/render_value.py
@classmethod
def retrieve_supported_render_combinations(cls) -> Iterable[Tuple[str, str]]:
result = []
for attr in dir(cls):
if (
len(attr) <= 16
or not attr.startswith("render__")
or "__as__" not in attr
):
continue
attr = attr[8:]
end_start_type = attr.find("__as__")
source_type = attr[0:end_start_type]
target_type = attr[end_start_type + 6 :] # noqa
result.append((source_type, target_type))
return result
ValueTypeRenderModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/render_value.py
class ValueTypeRenderModule(KiaraModule):
_module_type_name = "value.render"
_config_cls = RenderValueConfig
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
assert source_type not in ["target", "base_name"]
schema = {
source_type: {"type": source_type, "doc": "The value to render."},
"render_config": {
"type": "any",
"doc": "Value type dependent render configuration.",
"optional": True,
},
}
return schema
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"rendered_value": {
"type": self.get_config_value("target_type"),
"doc": "The rendered value.",
}
}
def process(self, inputs: ValueMap, outputs: ValueMap):
source_type = self.get_config_value("source_type")
target_type = self.get_config_value("target_type")
source_value = inputs.get_value_obj(source_type)
render_config = inputs.get_value_obj("render_config")
data_type_cls = source_value.data_type_class.get_class()
data_type = data_type_cls(**source_value.value_schema.type_config)
func_name = f"render_as__{target_type}"
func = getattr(data_type, func_name)
render_config_dict = render_config.data
if render_config_dict is None:
render_config_dict = {}
result = func(value=source_value, render_config=render_config_dict)
# TODO: check we have the correct type?
outputs.set_value("rendered_value", result)
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶Source code in kiara/modules/included_core_modules/render_value.py
class RenderValueConfig(KiaraModuleConfig):
source_type: str = Field(description="The value type of the source value.")
target_type: str = Field(description="The value type of the rendered value.")
@validator("source_type")
def validate_source_type(cls, value):
if value == "render_config":
raise ValueError(f"Invalid source type: {value}.")
return value
source_type: str
pydantic-field
required
¶The value type of the source value.
target_type: str
pydantic-field
required
¶The value type of the rendered value.
validate_source_type(value)
classmethod
¶Source code in kiara/modules/included_core_modules/render_value.py
@validator("source_type")
def validate_source_type(cls, value):
if value == "render_config":
raise ValueError(f"Invalid source type: {value}.")
return value
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/render_value.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
assert source_type not in ["target", "base_name"]
schema = {
source_type: {"type": source_type, "doc": "The value to render."},
"render_config": {
"type": "any",
"doc": "Value type dependent render configuration.",
"optional": True,
},
}
return schema
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/render_value.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"rendered_value": {
"type": self.get_config_value("target_type"),
"doc": "The rendered value.",
}
}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/render_value.py
def process(self, inputs: ValueMap, outputs: ValueMap):
source_type = self.get_config_value("source_type")
target_type = self.get_config_value("target_type")
source_value = inputs.get_value_obj(source_type)
render_config = inputs.get_value_obj("render_config")
data_type_cls = source_value.data_type_class.get_class()
data_type = data_type_cls(**source_value.value_schema.type_config)
func_name = f"render_as__{target_type}"
func = getattr(data_type, func_name)
render_config_dict = render_config.data
if render_config_dict is None:
render_config_dict = {}
result = func(value=source_value, render_config=render_config_dict)
# TODO: check we have the correct type?
outputs.set_value("rendered_value", result)
serialization
¶
PickleModule (SerializeValueModule)
¶Source code in kiara/modules/included_core_modules/serialization.py
class PickleModule(SerializeValueModule):
_module_type_name = "value.serialize.pickle"
def get_serialization_format_name(self):
return "pickle"
def from__any(self, value: Value, config: Dict[str, Any]):
"""Serialize any Python object into a bytes array using 'pickle'."""
import pickle5 as pickle
pickled = pickle.dumps(value.data, protocol=5)
data = {"value": pickled}
data_type_name = value.data_type_name
deserialization_config: Dict[str, Any] = {
"module_type": "value.serialize.pickle",
"module_config": {"target_type": data_type_name},
"output_name": data_type_name,
}
ser_val = SerializedValueModel(
data=data,
deserialization_config=DeserializationConfig.construct(
**deserialization_config
),
)
return ser_val
from__any(self, value, config)
¶Serialize any Python object into a bytes array using 'pickle'.
Source code in kiara/modules/included_core_modules/serialization.py
def from__any(self, value: Value, config: Dict[str, Any]):
"""Serialize any Python object into a bytes array using 'pickle'."""
import pickle5 as pickle
pickled = pickle.dumps(value.data, protocol=5)
data = {"value": pickled}
data_type_name = value.data_type_name
deserialization_config: Dict[str, Any] = {
"module_type": "value.serialize.pickle",
"module_config": {"target_type": data_type_name},
"output_name": data_type_name,
}
ser_val = SerializedValueModel(
data=data,
deserialization_config=DeserializationConfig.construct(
**deserialization_config
),
)
return ser_val
get_serialization_format_name(self)
¶Source code in kiara/modules/included_core_modules/serialization.py
def get_serialization_format_name(self):
return "pickle"
SerializeConfig (KiaraModuleConfig)
pydantic-model
¶Source code in kiara/modules/included_core_modules/serialization.py
class SerializeConfig(KiaraModuleConfig):
source_type: str = Field(description="The value type of the source.")
@validator("source_type")
def validate_source_type(cls, value):
if value == "serialization_config":
raise ValueError(f"Invalid source type: {value}.")
return value
source_type: str
pydantic-field
required
¶The value type of the source.
validate_source_type(value)
classmethod
¶Source code in kiara/modules/included_core_modules/serialization.py
@validator("source_type")
def validate_source_type(cls, value):
if value == "serialization_config":
raise ValueError(f"Invalid source type: {value}.")
return value
SerializeValueModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/serialization.py
class SerializeValueModule(KiaraModule):
_config_cls = SerializeConfig
@classmethod
def retrieve_supported_source_types(cls) -> Iterable[str]:
result = []
for attr in dir(cls):
if attr.startswith("from__"):
result.append(attr[6:])
return result
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
return {
source_type: {"type": source_type, "doc": "The value to serialize."},
"serialization_config": {
"type": "any",
"doc": "Serialization-format specific configuration.",
"optional": True,
},
}
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"serialized_value": {
"type": SERIALIZED_DATA_TYPE_NAME,
"type_config": {"format_name": self.get_serialization_format_name()},
"doc": "The value in serialized form.",
}
}
@abc.abstractmethod
def get_serialization_format_name(self) -> str:
pass
def process(self, inputs: ValueMap, outputs: ValueMap) -> None:
value = inputs.get_value_obj(self.get_config_value("source_type"))
config = inputs.get_value_obj("serialization_config")
func_name = f"from__{self.get_config_value('source_type')}"
func = getattr(self, func_name)
if config.is_set:
_config = config.data
else:
_config = {}
result: SerializedValueModel = func(value=value, config=_config)
outputs.set_value("serialized_value", result)
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶Source code in kiara/modules/included_core_modules/serialization.py
class SerializeConfig(KiaraModuleConfig):
source_type: str = Field(description="The value type of the source.")
@validator("source_type")
def validate_source_type(cls, value):
if value == "serialization_config":
raise ValueError(f"Invalid source type: {value}.")
return value
source_type: str
pydantic-field
required
¶The value type of the source.
validate_source_type(value)
classmethod
¶Source code in kiara/modules/included_core_modules/serialization.py
@validator("source_type")
def validate_source_type(cls, value):
if value == "serialization_config":
raise ValueError(f"Invalid source type: {value}.")
return value
create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/serialization.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
source_type = self.get_config_value("source_type")
return {
source_type: {"type": source_type, "doc": "The value to serialize."},
"serialization_config": {
"type": "any",
"doc": "Serialization-format specific configuration.",
"optional": True,
},
}
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/serialization.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
return {
"serialized_value": {
"type": SERIALIZED_DATA_TYPE_NAME,
"type_config": {"format_name": self.get_serialization_format_name()},
"doc": "The value in serialized form.",
}
}
get_serialization_format_name(self)
¶Source code in kiara/modules/included_core_modules/serialization.py
@abc.abstractmethod
def get_serialization_format_name(self) -> str:
pass
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/serialization.py
def process(self, inputs: ValueMap, outputs: ValueMap) -> None:
value = inputs.get_value_obj(self.get_config_value("source_type"))
config = inputs.get_value_obj("serialization_config")
func_name = f"from__{self.get_config_value('source_type')}"
func = getattr(self, func_name)
if config.is_set:
_config = config.data
else:
_config = {}
result: SerializedValueModel = func(value=value, config=_config)
outputs.set_value("serialized_value", result)
retrieve_supported_source_types()
classmethod
¶Source code in kiara/modules/included_core_modules/serialization.py
@classmethod
def retrieve_supported_source_types(cls) -> Iterable[str]:
result = []
for attr in dir(cls):
if attr.startswith("from__"):
result.append(attr[6:])
return result
UnpickleConfig (KiaraModuleConfig)
pydantic-model
¶
UnpickleModule (KiaraModule)
¶Source code in kiara/modules/included_core_modules/serialization.py
class UnpickleModule(KiaraModule):
_module_type_name = "value.serialize.unpickle"
_config_cls = UnpickleConfig
def _retrieve_module_characteristics(self) -> ModuleCharacteristics:
return ModuleCharacteristics(is_internal=True)
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
target_type = self.get_config_value("target_type")
return {
"bytes": {
"type": "bytes",
"doc": f"The serialized bytes of the '{target_type}' value.",
}
}
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
target_type = self.get_config_value("target_type")
return {
target_type: {
"type": target_type,
"doc": "The type of the value to unpickle.",
}
}
def process(self, inputs: ValueMap, outputs: ValueMap):
import pickle5 as pickle
target_type = self.get_config_value("target_type")
_bytes = inputs.get_value_data("bytes")
data = pickle.loads(_bytes)
outputs.set_value(target_type, data)
_config_cls (KiaraModuleConfig)
private
pydantic-model
¶create_inputs_schema(self)
¶Return the schema for this types' inputs.
Source code in kiara/modules/included_core_modules/serialization.py
def create_inputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
target_type = self.get_config_value("target_type")
return {
"bytes": {
"type": "bytes",
"doc": f"The serialized bytes of the '{target_type}' value.",
}
}
create_outputs_schema(self)
¶Return the schema for this types' outputs.
Source code in kiara/modules/included_core_modules/serialization.py
def create_outputs_schema(
self,
) -> Mapping[str, Union[ValueSchema, Mapping[str, Any]]]:
target_type = self.get_config_value("target_type")
return {
target_type: {
"type": target_type,
"doc": "The type of the value to unpickle.",
}
}
process(self, inputs, outputs)
¶Source code in kiara/modules/included_core_modules/serialization.py
def process(self, inputs: ValueMap, outputs: ValueMap):
import pickle5 as pickle
target_type = self.get_config_value("target_type")
_bytes = inputs.get_value_data("bytes")
data = pickle.loads(_bytes)
outputs.set_value(target_type, data)
operations
special
¶
OPERATION_TYPE_DETAILS
¶
Classes¶
OperationType (ABC, Generic)
¶
Source code in kiara/operations/__init__.py
class OperationType(abc.ABC, Generic[OPERATION_TYPE_DETAILS]):
def __init__(self, kiara: "Kiara", op_type_name: str):
self._kiara: Kiara = kiara
self._op_type_name: str = op_type_name
@property
def operations(self) -> Mapping[str, Operation]:
return {
op_id: self._kiara.operation_registry.get_operation(op_id)
for op_id in self._kiara.operation_registry.operations_by_type[
self._op_type_name
]
}
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
return []
@abc.abstractmethod
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[OPERATION_TYPE_DETAILS]:
"""Check whether the provided module is a valid operation for this type."""
def retrieve_operation_details(
self, operation: Union[Operation, str]
) -> OPERATION_TYPE_DETAILS:
"""Retrieve operation details for provided operation.
This is really just a utility method, to make the type checker happy.
"""
if isinstance(operation, str):
operation = self.operations[operation]
return operation.operation_details # type: ignore
operations: Mapping[str, kiara.models.module.operation.Operation]
property
readonly
¶Methods¶
check_matching_operation(self, module)
¶Check whether the provided module is a valid operation for this type.
Source code in kiara/operations/__init__.py
@abc.abstractmethod
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[OPERATION_TYPE_DETAILS]:
"""Check whether the provided module is a valid operation for this type."""
retrieve_included_operation_configs(self)
¶Source code in kiara/operations/__init__.py
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
return []
retrieve_operation_details(self, operation)
¶Retrieve operation details for provided operation.
This is really just a utility method, to make the type checker happy.
Source code in kiara/operations/__init__.py
def retrieve_operation_details(
self, operation: Union[Operation, str]
) -> OPERATION_TYPE_DETAILS:
"""Retrieve operation details for provided operation.
This is really just a utility method, to make the type checker happy.
"""
if isinstance(operation, str):
operation = self.operations[operation]
return operation.operation_details # type: ignore
Modules¶
included_core_operations
special
¶
logger
¶Classes¶
CustomModuleOperationDetails (OperationDetails)
pydantic-model
¶Source code in kiara/operations/included_core_operations/__init__.py
class CustomModuleOperationDetails(OperationDetails):
@classmethod
def create_from_module(cls, module: KiaraModule):
return CustomModuleOperationDetails(
operation_id=module.module_type_name,
module_inputs_schema=module.inputs_schema,
module_outputs_schema=module.outputs_schema,
)
module_inputs_schema: Mapping[str, ValueSchema] = Field(
description="The input schemas of the module."
)
module_outputs_schema: Mapping[str, ValueSchema] = Field(
description="The output schemas of the module."
)
_op_schema: OperationSchema = PrivateAttr(default=None)
def get_operation_schema(self) -> OperationSchema:
if self._op_schema is not None:
return self._op_schema
self._op_schema = OperationSchema(
alias=self.operation_id,
inputs_schema=self.module_inputs_schema,
outputs_schema=self.module_outputs_schema,
)
return self._op_schema
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return inputs
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return outputs
module_inputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
pydantic-field
required
¶The input schemas of the module.
module_outputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
pydantic-field
required
¶The output schemas of the module.
create_from_module(module)
classmethod
¶Source code in kiara/operations/included_core_operations/__init__.py
@classmethod
def create_from_module(cls, module: KiaraModule):
return CustomModuleOperationDetails(
operation_id=module.module_type_name,
module_inputs_schema=module.inputs_schema,
module_outputs_schema=module.outputs_schema,
)
create_module_inputs(self, inputs)
¶Source code in kiara/operations/included_core_operations/__init__.py
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return inputs
create_operation_outputs(self, outputs)
¶Source code in kiara/operations/included_core_operations/__init__.py
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return outputs
get_operation_schema(self)
¶Source code in kiara/operations/included_core_operations/__init__.py
def get_operation_schema(self) -> OperationSchema:
if self._op_schema is not None:
return self._op_schema
self._op_schema = OperationSchema(
alias=self.operation_id,
inputs_schema=self.module_inputs_schema,
outputs_schema=self.module_outputs_schema,
)
return self._op_schema
CustomModuleOperationType (OperationType)
¶Source code in kiara/operations/included_core_operations/__init__.py
class CustomModuleOperationType(OperationType[CustomModuleOperationDetails]):
_operation_type_name = "custom_module"
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = []
for name, module_cls in self._kiara.module_type_classes.items():
mod_conf = module_cls._config_cls
if mod_conf.requires_config():
logger.debug(
"ignore.custom_operation",
module_type=name,
reason="config required",
)
continue
doc = DocumentationMetadataModel.from_class_doc(module_cls)
oc = ManifestOperationConfig(module_type=name, doc=doc)
result.append(oc)
return result
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[CustomModuleOperationDetails]:
mod_conf = module.__class__._config_cls
if not mod_conf.requires_config():
is_internal = module.characteristics.is_internal
op_details = CustomModuleOperationDetails.create_operation_details(
operation_id=module.module_type_name,
module_inputs_schema=module.inputs_schema,
module_outputs_schema=module.outputs_schema,
is_internal_operation=is_internal,
)
return op_details
else:
return None
check_matching_operation(self, module)
¶Check whether the provided module is a valid operation for this type.
Source code in kiara/operations/included_core_operations/__init__.py
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[CustomModuleOperationDetails]:
mod_conf = module.__class__._config_cls
if not mod_conf.requires_config():
is_internal = module.characteristics.is_internal
op_details = CustomModuleOperationDetails.create_operation_details(
operation_id=module.module_type_name,
module_inputs_schema=module.inputs_schema,
module_outputs_schema=module.outputs_schema,
is_internal_operation=is_internal,
)
return op_details
else:
return None
retrieve_included_operation_configs(self)
¶Source code in kiara/operations/included_core_operations/__init__.py
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = []
for name, module_cls in self._kiara.module_type_classes.items():
mod_conf = module_cls._config_cls
if mod_conf.requires_config():
logger.debug(
"ignore.custom_operation",
module_type=name,
reason="config required",
)
continue
doc = DocumentationMetadataModel.from_class_doc(module_cls)
oc = ManifestOperationConfig(module_type=name, doc=doc)
result.append(oc)
return result
Modules¶
create_from
¶logger
¶
CreateFromOperationType (OperationType)
¶Source code in kiara/operations/included_core_operations/create_from.py
class CreateFromOperationType(OperationType[CreateValueFromDetails]):
_operation_type_name = "create_from"
def _calculate_op_id(self, source_type: str, target_type: str):
if source_type == "any":
operation_id = f"create.{target_type}"
else:
operation_id = f"create.{target_type}.from.{source_type}"
return operation_id
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = {}
for name, module_cls in self._kiara.module_type_classes.items():
if not hasattr(module_cls, "retrieve_supported_create_combinations"):
continue
try:
supported_combinations = module_cls.retrieve_supported_create_combinations() # type: ignore
for sup_comb in supported_combinations:
source_type = sup_comb["source_type"]
target_type = sup_comb["target_type"]
func = sup_comb["func"]
if source_type not in self._kiara.data_type_names:
logger.debug(
"ignore.operation_config",
module_type=name,
reason=f"Source type '{source_type}' not registered.",
)
continue
if target_type not in self._kiara.data_type_names:
logger.debug(
"ignore.operation_config",
module_type=name,
reason=f"Target type '{target_type}' not registered.",
)
continue
if not hasattr(module_cls, func):
logger.debug(
"ignore.operation_config",
module_type=name,
reason=f"Specified create function '{func}' not available.",
)
continue
mc = {"source_type": source_type, "target_type": target_type}
# TODO: check whether module config actually supports those, for now, only 'CreateFromModule' subtypes are supported
_func = getattr(module_cls, func)
doc = DocumentationMetadataModel.from_function(_func)
oc = ManifestOperationConfig(
module_type=name, module_config=mc, doc=doc
)
op_id = self._calculate_op_id(
source_type=source_type, target_type=target_type
)
result[op_id] = oc
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
logger.debug(
"ignore.create_operation_instance", module_type=name, reason=e
)
continue
return result.values()
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[CreateValueFromDetails]:
source_type = None
for field_name, schema in module.inputs_schema.items():
if field_name == schema.type:
if source_type is not None:
logger.debug(
"ignore.operation",
operation_type="create_from",
reason=f"more than one possible target type field: {field_name}",
)
return None
source_type = field_name
if source_type is None:
return None
target_type = None
for field_name, schema in module.outputs_schema.items():
if field_name == schema.type:
if target_type is not None:
logger.debug(
"ignore.operation",
operation_type="create_from",
reason=f"more than one possible target type field: {field_name}",
)
return None
target_type = field_name
if target_type is None:
return None
op_id = self._calculate_op_id(source_type=source_type, target_type=target_type)
if (
"any" in self._kiara.type_registry.get_type_lineage(target_type)
and target_type != "any"
):
is_internal = False
else:
is_internal = True
details = {
"operation_id": op_id,
"source_type": source_type,
"target_type": target_type,
"is_internal_operation": is_internal,
}
result = CreateValueFromDetails.create_operation_details(**details)
return result
check_matching_operation(self, module)
¶Check whether the provided module is a valid operation for this type.
Source code in kiara/operations/included_core_operations/create_from.py
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[CreateValueFromDetails]:
source_type = None
for field_name, schema in module.inputs_schema.items():
if field_name == schema.type:
if source_type is not None:
logger.debug(
"ignore.operation",
operation_type="create_from",
reason=f"more than one possible target type field: {field_name}",
)
return None
source_type = field_name
if source_type is None:
return None
target_type = None
for field_name, schema in module.outputs_schema.items():
if field_name == schema.type:
if target_type is not None:
logger.debug(
"ignore.operation",
operation_type="create_from",
reason=f"more than one possible target type field: {field_name}",
)
return None
target_type = field_name
if target_type is None:
return None
op_id = self._calculate_op_id(source_type=source_type, target_type=target_type)
if (
"any" in self._kiara.type_registry.get_type_lineage(target_type)
and target_type != "any"
):
is_internal = False
else:
is_internal = True
details = {
"operation_id": op_id,
"source_type": source_type,
"target_type": target_type,
"is_internal_operation": is_internal,
}
result = CreateValueFromDetails.create_operation_details(**details)
return result
retrieve_included_operation_configs(self)
¶Source code in kiara/operations/included_core_operations/create_from.py
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = {}
for name, module_cls in self._kiara.module_type_classes.items():
if not hasattr(module_cls, "retrieve_supported_create_combinations"):
continue
try:
supported_combinations = module_cls.retrieve_supported_create_combinations() # type: ignore
for sup_comb in supported_combinations:
source_type = sup_comb["source_type"]
target_type = sup_comb["target_type"]
func = sup_comb["func"]
if source_type not in self._kiara.data_type_names:
logger.debug(
"ignore.operation_config",
module_type=name,
reason=f"Source type '{source_type}' not registered.",
)
continue
if target_type not in self._kiara.data_type_names:
logger.debug(
"ignore.operation_config",
module_type=name,
reason=f"Target type '{target_type}' not registered.",
)
continue
if not hasattr(module_cls, func):
logger.debug(
"ignore.operation_config",
module_type=name,
reason=f"Specified create function '{func}' not available.",
)
continue
mc = {"source_type": source_type, "target_type": target_type}
# TODO: check whether module config actually supports those, for now, only 'CreateFromModule' subtypes are supported
_func = getattr(module_cls, func)
doc = DocumentationMetadataModel.from_function(_func)
oc = ManifestOperationConfig(
module_type=name, module_config=mc, doc=doc
)
op_id = self._calculate_op_id(
source_type=source_type, target_type=target_type
)
result[op_id] = oc
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
logger.debug(
"ignore.create_operation_instance", module_type=name, reason=e
)
continue
return result.values()
CreateValueFromDetails (BaseOperationDetails)
pydantic-model
¶Source code in kiara/operations/included_core_operations/create_from.py
class CreateValueFromDetails(BaseOperationDetails):
source_type: str = Field(description="The type of the value to be created.")
target_type: str = Field(description="The result type.")
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {
self.source_type: {"type": self.source_type, "doc": "The source value."},
}
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {
self.target_type: {"type": self.target_type, "doc": "The result value."}
}
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return inputs
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return outputs
source_type: str
pydantic-field
required
¶The type of the value to be created.
target_type: str
pydantic-field
required
¶The result type.
create_module_inputs(self, inputs)
¶Source code in kiara/operations/included_core_operations/create_from.py
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return inputs
create_operation_outputs(self, outputs)
¶Source code in kiara/operations/included_core_operations/create_from.py
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return outputs
retrieve_inputs_schema(self)
¶Source code in kiara/operations/included_core_operations/create_from.py
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {
self.source_type: {"type": self.source_type, "doc": "The source value."},
}
retrieve_outputs_schema(self)
¶Source code in kiara/operations/included_core_operations/create_from.py
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {
self.target_type: {"type": self.target_type, "doc": "The result value."}
}
metadata
¶
ExtractMetadataDetails (BaseOperationDetails)
pydantic-model
¶A model that contains information needed to describe an 'extract_metadata' operation.
Source code in kiara/operations/included_core_operations/metadata.py
class ExtractMetadataDetails(BaseOperationDetails):
"""A model that contains information needed to describe an 'extract_metadata' operation."""
data_type: str = Field(
description="The data type this metadata operation can be used with."
)
metadata_key: str = Field(description="The metadata key.")
input_field_name: str = Field(description="The input field name.")
result_field_name: str = Field(description="The result field name.")
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {
"value": {
"type": self.data_type,
"doc": f"The {self.data_type} value to extract metadata from.",
}
}
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {"value_metadata": {"type": "value_metadata", "doc": "The metadata."}}
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return {self.input_field_name: inputs["value"]}
def create_operation_outputs(self, outputs: ValueMap) -> ValueMap:
return outputs
data_type: str
pydantic-field
required
¶The data type this metadata operation can be used with.
input_field_name: str
pydantic-field
required
¶The input field name.
metadata_key: str
pydantic-field
required
¶The metadata key.
result_field_name: str
pydantic-field
required
¶The result field name.
create_module_inputs(self, inputs)
¶Source code in kiara/operations/included_core_operations/metadata.py
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return {self.input_field_name: inputs["value"]}
create_operation_outputs(self, outputs)
¶Source code in kiara/operations/included_core_operations/metadata.py
def create_operation_outputs(self, outputs: ValueMap) -> ValueMap:
return outputs
retrieve_inputs_schema(self)
¶Source code in kiara/operations/included_core_operations/metadata.py
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {
"value": {
"type": self.data_type,
"doc": f"The {self.data_type} value to extract metadata from.",
}
}
retrieve_outputs_schema(self)
¶Source code in kiara/operations/included_core_operations/metadata.py
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {"value_metadata": {"type": "value_metadata", "doc": "The metadata."}}
ExtractMetadataOperationType (OperationType)
¶An operation that extracts metadata of a specific type from value data.
For a module profile to be picked up by this operation type, it needs to have: - exactly one input field - that input field must have the same name as its value type, or be 'value' - exactly one output field, whose field name is called 'value_metadata', and where the value has the type 'internal_model'
Source code in kiara/operations/included_core_operations/metadata.py
class ExtractMetadataOperationType(OperationType[ExtractMetadataDetails]):
"""An operation that extracts metadata of a specific type from value data.
For a module profile to be picked up by this operation type, it needs to have:
- exactly one input field
- that input field must have the same name as its value type, or be 'value'
- exactly one output field, whose field name is called 'value_metadata', and where the value has the type 'internal_model'
"""
_operation_type_name = "extract_metadata"
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
all_models = find_all_value_metadata_models()
result = []
for metadata_key, model_cls in all_models.items():
data_types = model_cls.retrieve_supported_data_types()
if isinstance(data_types, str):
data_types = [data_types]
for data_type in data_types:
config = {
"module_type": "value.extract_metadata",
"module_config": {
"data_type": data_type,
"metadata_model": PythonClass.from_class(model_cls),
},
"doc": f"Extract '{metadata_key}' metadata for value type '{data_type}'.",
}
result.append(config)
return result
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[ExtractMetadataDetails]:
if len(module.outputs_schema) != 1:
return None
if (
"value_metadata" not in module.outputs_schema
or module.outputs_schema["value_metadata"].type != "internal_model"
):
return None
if len(module.inputs_schema) != 1:
return None
input_field_name = next(iter(module.inputs_schema.keys()))
input_schema = module.inputs_schema.get(input_field_name)
assert input_schema is not None
if input_field_name != input_schema.type and input_field_name != "value":
return None
data_type_name = module.inputs_schema["value"].type
# metadata_key=module.get_config_value("metadata_key")
metadata_model: PythonClass = module.get_config_value("metadata_model")
metadata_key = metadata_model.get_class()._metadata_key # type: ignore
if data_type_name == "any":
op_id = f"extract.{metadata_key}.metadata"
else:
op_id = f"extract.{metadata_key}.metadata.from.{data_type_name}"
details = ExtractMetadataDetails.create_operation_details(
operation_id=op_id,
data_type=data_type_name,
metadata_key=metadata_key,
input_field_name=input_field_name,
result_field_name="value_metadata",
is_internal_operation=True,
)
return details
def get_operations_for_data_type(self, data_type: str) -> Mapping[str, Operation]:
"""Return all available metadata extract operations for the provided type (and it's parent types).
Arguments:
data_type: the value type
Returns:
a mapping with the metadata type as key, and the operation as value
"""
lineage = set(
self._kiara.type_registry.get_type_lineage(data_type_name=data_type)
)
result = {}
for op_id, op in self.operations.items():
op_details = self.retrieve_operation_details(op)
included = op_details.data_type in lineage
if not included:
continue
metadata_key = op_details.metadata_key
if metadata_key in result:
raise Exception(
f"Duplicate metadata operations for type '{metadata_key}'."
)
result[metadata_key] = op
return result
check_matching_operation(self, module)
¶Check whether the provided module is a valid operation for this type.
Source code in kiara/operations/included_core_operations/metadata.py
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[ExtractMetadataDetails]:
if len(module.outputs_schema) != 1:
return None
if (
"value_metadata" not in module.outputs_schema
or module.outputs_schema["value_metadata"].type != "internal_model"
):
return None
if len(module.inputs_schema) != 1:
return None
input_field_name = next(iter(module.inputs_schema.keys()))
input_schema = module.inputs_schema.get(input_field_name)
assert input_schema is not None
if input_field_name != input_schema.type and input_field_name != "value":
return None
data_type_name = module.inputs_schema["value"].type
# metadata_key=module.get_config_value("metadata_key")
metadata_model: PythonClass = module.get_config_value("metadata_model")
metadata_key = metadata_model.get_class()._metadata_key # type: ignore
if data_type_name == "any":
op_id = f"extract.{metadata_key}.metadata"
else:
op_id = f"extract.{metadata_key}.metadata.from.{data_type_name}"
details = ExtractMetadataDetails.create_operation_details(
operation_id=op_id,
data_type=data_type_name,
metadata_key=metadata_key,
input_field_name=input_field_name,
result_field_name="value_metadata",
is_internal_operation=True,
)
return details
get_operations_for_data_type(self, data_type)
¶Return all available metadata extract operations for the provided type (and it's parent types).
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
data_type |
str |
the value type |
required |
Returns:
| Type | Description |
|---|---|
Mapping[str, kiara.models.module.operation.Operation] |
a mapping with the metadata type as key, and the operation as value |
Source code in kiara/operations/included_core_operations/metadata.py
def get_operations_for_data_type(self, data_type: str) -> Mapping[str, Operation]:
"""Return all available metadata extract operations for the provided type (and it's parent types).
Arguments:
data_type: the value type
Returns:
a mapping with the metadata type as key, and the operation as value
"""
lineage = set(
self._kiara.type_registry.get_type_lineage(data_type_name=data_type)
)
result = {}
for op_id, op in self.operations.items():
op_details = self.retrieve_operation_details(op)
included = op_details.data_type in lineage
if not included:
continue
metadata_key = op_details.metadata_key
if metadata_key in result:
raise Exception(
f"Duplicate metadata operations for type '{metadata_key}'."
)
result[metadata_key] = op
return result
retrieve_included_operation_configs(self)
¶Source code in kiara/operations/included_core_operations/metadata.py
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
all_models = find_all_value_metadata_models()
result = []
for metadata_key, model_cls in all_models.items():
data_types = model_cls.retrieve_supported_data_types()
if isinstance(data_types, str):
data_types = [data_types]
for data_type in data_types:
config = {
"module_type": "value.extract_metadata",
"module_config": {
"data_type": data_type,
"metadata_model": PythonClass.from_class(model_cls),
},
"doc": f"Extract '{metadata_key}' metadata for value type '{data_type}'.",
}
result.append(config)
return result
persistence
¶
PersistValueDetails (BaseOperationDetails)
pydantic-model
¶Source code in kiara/operations/included_core_operations/persistence.py
class PersistValueDetails(BaseOperationDetails):
value_input_field: str = Field(
description="The (input) field name containing the value to be persisted."
)
value_input_type: str = Field(description="The type of the value to be persisted.")
load_config_output_field: str = Field(
description="The (output) field name containing the details of the persisted value."
)
persistence_target: str = Field(description="The name of the persistence target.")
persistence_format: str = Field(description="The persistence format.")
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {
"value": {"type": "any", "doc": "The value to persist."},
"persitence_config": {
"type": "any",
"doc": "(Optional) configuration for the persistance process.",
},
}
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {
"load_config": {"type": "load_config", "doc": "The saved value details."},
"bytes_structure": {
"type": "any",
"doc": "A structure of serialized bytes.",
},
}
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return {
self.value_input_field: inputs["value"],
"persistence_config": inputs.get("persistence_config", {}),
}
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return outputs
load_config_output_field: str
pydantic-field
required
¶The (output) field name containing the details of the persisted value.
persistence_format: str
pydantic-field
required
¶The persistence format.
persistence_target: str
pydantic-field
required
¶The name of the persistence target.
value_input_field: str
pydantic-field
required
¶The (input) field name containing the value to be persisted.
value_input_type: str
pydantic-field
required
¶The type of the value to be persisted.
create_module_inputs(self, inputs)
¶Source code in kiara/operations/included_core_operations/persistence.py
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return {
self.value_input_field: inputs["value"],
"persistence_config": inputs.get("persistence_config", {}),
}
create_operation_outputs(self, outputs)
¶Source code in kiara/operations/included_core_operations/persistence.py
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return outputs
retrieve_inputs_schema(self)
¶Source code in kiara/operations/included_core_operations/persistence.py
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {
"value": {"type": "any", "doc": "The value to persist."},
"persitence_config": {
"type": "any",
"doc": "(Optional) configuration for the persistance process.",
},
}
retrieve_outputs_schema(self)
¶Source code in kiara/operations/included_core_operations/persistence.py
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {
"load_config": {"type": "load_config", "doc": "The saved value details."},
"bytes_structure": {
"type": "any",
"doc": "A structure of serialized bytes.",
},
}
PersistValueOperationType (OperationType)
¶An operation that takes a value, and saves it to disk and returns details about how to re-assemble the value (via a [kiara.data_types.included_core_types.persistence.LoadConfigValueType] object).
For a module profile to be picked up by this operation type, it needs to:
- exactly one output field of type load_config
- either one of (in this order):
- exactly one input field
- one input field where the field name equals the type name
- an input field called 'value'
Source code in kiara/operations/included_core_operations/persistence.py
class PersistValueOperationType(OperationType[PersistValueDetails]):
"""An operation that takes a value, and saves it to disk and returns details about how to re-assemble the value (via a [kiara.data_types.included_core_types.persistence.LoadConfigValueType] object).
For a module profile to be picked up by this operation type, it needs to:
- exactly one output field of type `load_config`
- either one of (in this order):
- exactly one input field
- one input field where the field name equals the type name
- an input field called 'value'
"""
_operation_type_name = "persist_value"
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = []
for name, module_cls in self._kiara.module_type_classes.items():
if not issubclass(module_cls, PersistValueModule):
continue
for st in module_cls.retrieve_supported_source_types():
func_name = f"data_type__{st}"
attr = getattr(module_cls, func_name)
doc = DocumentationMetadataModel.from_function(attr)
mc = {"source_type": st}
oc = ManifestOperationConfig(
module_type=name, module_config=mc, doc=doc
)
result.append(oc)
return result
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[PersistValueDetails]:
details = self.extract_details(module)
if details is None:
return None
else:
return details
def extract_details(self, module: "KiaraModule") -> Optional[PersistValueDetails]:
if len(module.inputs_schema) != 1:
return None
# if "save" not in module.module_type_name:
# return None
match = None
for field_name, schema in module.outputs_schema.items():
if schema.type != LOAD_CONFIG_DATA_TYPE_NAME:
continue
else:
if match is not None:
log_message(
"ignore.operation",
reason=f"More than one field of type '{LOAD_CONFIG_DATA_TYPE_NAME}'",
module_type=module.module_type_name,
)
continue
else:
match = field_name
if not match:
return None
input_field_match = None
for field_name, schema in module.inputs_schema.items():
if field_name == schema.type:
if input_field_match is not None:
# we can't deal (yet) with multiple fields
log_message(
"operation.ignore",
module=module.module_type_name,
reason=f"more than one input fields of type '{schema.type}'",
)
input_field_match = None
break
else:
input_field_match = field_name
if input_field_match is not None:
input_field = input_field_match
else:
return None
input_field_type = module.inputs_schema[input_field].type
value_schema: ValueSchema = module.outputs_schema[match]
load_config_type: LoadConfigValueType = self._kiara.type_registry.retrieve_data_type( # type: ignore # type: ignore
data_type_name=value_schema.type,
data_type_config=value_schema.type_config,
) # type: ignore
persistence_target = load_config_type.type_config.persistence_target # type: ignore
persistence_format = load_config_type.type_config.persistence_format # type: ignore
if input_field_type == "any":
operation_id = f"save.to.{persistence_target}.as.{persistence_format}"
else:
operation_id = f"save.{input_field_type}.to.{persistence_target}.as.{persistence_format}"
details = {
"operation_id": operation_id,
"value_input_field": input_field,
"value_input_type": input_field_type,
"load_config_output_field": match,
"persistence_target": persistence_target,
"persistence_format": persistence_format,
"is_internal_operation": True,
}
result = PersistValueDetails.create_operation_details(**details)
return result
def get_operation_for_data_type(self, type_name: str) -> Operation:
lineage = self._kiara.type_registry.get_type_lineage(type_name)
persist_op: Optional[Operation] = None
for data_type in lineage:
match = []
for op in self.operations.values():
details = self.retrieve_operation_details(op)
if details.value_input_type == data_type:
match.append(op)
if match:
if len(match) > 1:
raise Exception(
f"Multiple serialization operations found for value type '{type_name}'. This is not supported (yet)."
)
persist_op = match[0]
break
if persist_op is None:
raise Exception(f"Can't find persist operation for type '{type_name}'.")
return persist_op
check_matching_operation(self, module)
¶Check whether the provided module is a valid operation for this type.
Source code in kiara/operations/included_core_operations/persistence.py
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[PersistValueDetails]:
details = self.extract_details(module)
if details is None:
return None
else:
return details
extract_details(self, module)
¶Source code in kiara/operations/included_core_operations/persistence.py
def extract_details(self, module: "KiaraModule") -> Optional[PersistValueDetails]:
if len(module.inputs_schema) != 1:
return None
# if "save" not in module.module_type_name:
# return None
match = None
for field_name, schema in module.outputs_schema.items():
if schema.type != LOAD_CONFIG_DATA_TYPE_NAME:
continue
else:
if match is not None:
log_message(
"ignore.operation",
reason=f"More than one field of type '{LOAD_CONFIG_DATA_TYPE_NAME}'",
module_type=module.module_type_name,
)
continue
else:
match = field_name
if not match:
return None
input_field_match = None
for field_name, schema in module.inputs_schema.items():
if field_name == schema.type:
if input_field_match is not None:
# we can't deal (yet) with multiple fields
log_message(
"operation.ignore",
module=module.module_type_name,
reason=f"more than one input fields of type '{schema.type}'",
)
input_field_match = None
break
else:
input_field_match = field_name
if input_field_match is not None:
input_field = input_field_match
else:
return None
input_field_type = module.inputs_schema[input_field].type
value_schema: ValueSchema = module.outputs_schema[match]
load_config_type: LoadConfigValueType = self._kiara.type_registry.retrieve_data_type( # type: ignore # type: ignore
data_type_name=value_schema.type,
data_type_config=value_schema.type_config,
) # type: ignore
persistence_target = load_config_type.type_config.persistence_target # type: ignore
persistence_format = load_config_type.type_config.persistence_format # type: ignore
if input_field_type == "any":
operation_id = f"save.to.{persistence_target}.as.{persistence_format}"
else:
operation_id = f"save.{input_field_type}.to.{persistence_target}.as.{persistence_format}"
details = {
"operation_id": operation_id,
"value_input_field": input_field,
"value_input_type": input_field_type,
"load_config_output_field": match,
"persistence_target": persistence_target,
"persistence_format": persistence_format,
"is_internal_operation": True,
}
result = PersistValueDetails.create_operation_details(**details)
return result
get_operation_for_data_type(self, type_name)
¶Source code in kiara/operations/included_core_operations/persistence.py
def get_operation_for_data_type(self, type_name: str) -> Operation:
lineage = self._kiara.type_registry.get_type_lineage(type_name)
persist_op: Optional[Operation] = None
for data_type in lineage:
match = []
for op in self.operations.values():
details = self.retrieve_operation_details(op)
if details.value_input_type == data_type:
match.append(op)
if match:
if len(match) > 1:
raise Exception(
f"Multiple serialization operations found for value type '{type_name}'. This is not supported (yet)."
)
persist_op = match[0]
break
if persist_op is None:
raise Exception(f"Can't find persist operation for type '{type_name}'.")
return persist_op
retrieve_included_operation_configs(self)
¶Source code in kiara/operations/included_core_operations/persistence.py
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = []
for name, module_cls in self._kiara.module_type_classes.items():
if not issubclass(module_cls, PersistValueModule):
continue
for st in module_cls.retrieve_supported_source_types():
func_name = f"data_type__{st}"
attr = getattr(module_cls, func_name)
doc = DocumentationMetadataModel.from_function(attr)
mc = {"source_type": st}
oc = ManifestOperationConfig(
module_type=name, module_config=mc, doc=doc
)
result.append(oc)
return result
pipeline
¶logger
¶
PipelineOperationDetails (OperationDetails)
pydantic-model
¶Source code in kiara/operations/included_core_operations/pipeline.py
class PipelineOperationDetails(OperationDetails):
# @classmethod
# def create_from_module(cls, module: KiaraModule):
#
# return PipelineOperationDetails(
# operation_id=module.module_type_name,
# pipeline_inputs_schema=module.inputs_schema,
# pipeline_outputs_schema=module.outputs_schema,
# )
pipeline_inputs_schema: Mapping[str, ValueSchema] = Field(
description="The input schema for the pipeline."
)
pipeline_outputs_schema: Mapping[str, ValueSchema] = Field(
description="The output schema for the pipeline."
)
_op_schema: OperationSchema = PrivateAttr(default=None)
def get_operation_schema(self) -> OperationSchema:
if self._op_schema is not None:
return self._op_schema
self._op_schema = OperationSchema(
alias=self.operation_id,
inputs_schema=self.pipeline_inputs_schema,
outputs_schema=self.pipeline_outputs_schema,
)
return self._op_schema
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return inputs
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return outputs
pipeline_inputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
pydantic-field
required
¶The input schema for the pipeline.
pipeline_outputs_schema: Mapping[str, kiara.models.values.value_schema.ValueSchema]
pydantic-field
required
¶The output schema for the pipeline.
create_module_inputs(self, inputs)
¶Source code in kiara/operations/included_core_operations/pipeline.py
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return inputs
create_operation_outputs(self, outputs)
¶Source code in kiara/operations/included_core_operations/pipeline.py
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return outputs
get_operation_schema(self)
¶Source code in kiara/operations/included_core_operations/pipeline.py
def get_operation_schema(self) -> OperationSchema:
if self._op_schema is not None:
return self._op_schema
self._op_schema = OperationSchema(
alias=self.operation_id,
inputs_schema=self.pipeline_inputs_schema,
outputs_schema=self.pipeline_outputs_schema,
)
return self._op_schema
PipelineOperationType (OperationType)
¶Source code in kiara/operations/included_core_operations/pipeline.py
class PipelineOperationType(OperationType[PipelineOperationDetails]):
_operation_type_name = "pipeline"
def __init__(self, kiara: "Kiara", op_type_name: str):
super().__init__(kiara=kiara, op_type_name=op_type_name)
self._pipelines = None
@property
def pipeline_data(self):
if self._pipelines is not None:
return self._pipelines
ignore_errors = False
pipeline_paths: List[str] = find_all_kiara_pipeline_paths(
skip_errors=ignore_errors
)
all_pipelines = []
for path in pipeline_paths:
if isinstance(path, str):
path = Path(os.path.expanduser(path))
elif isinstance(path, Iterable):
raise TypeError(f"Invalid type for path: {path}")
if not path.exists():
logger.warning(
"ignore.pipeline_path", path=path, reason="path does not exist"
)
continue
elif path.is_dir():
for root, dirnames, filenames in os.walk(path, topdown=True):
dirnames[:] = [d for d in dirnames if d not in DEFAULT_EXCLUDE_DIRS]
for filename in [
f
for f in filenames
if os.path.isfile(os.path.join(root, f))
and any(
f.endswith(ext) for ext in VALID_PIPELINE_FILE_EXTENSIONS
)
]:
full_path = os.path.join(root, filename)
try:
data = get_pipeline_details_from_path(path=full_path)
data = check_doc_sidecar(full_path, data)
# rel_path = os.path.relpath(os.path.dirname(full_path), path)
# if not rel_path or rel_path == ".":
# raise NotImplementedError()
# ns_name = name
# else:
# _rel_path = rel_path.replace(os.path.sep, ".")
# ns_name = f"{_rel_path}.{name}"
#
# if not ns_name:
# raise Exception(
# f"Could not determine namespace for pipeline file '{filename}'."
# )
# if ns_name in files.keys():
# raise Exception(
# f"Duplicate workflow name: {ns_name}"
# )
all_pipelines.append(data)
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
logger.warning(
"ignore.pipeline_file", path=full_path, reason=str(e)
)
elif path.is_file():
data = get_pipeline_details_from_path(path=path)
data = check_doc_sidecar(path, data)
all_pipelines.append(data)
pipelines = {}
for pipeline in all_pipelines:
name = pipeline["data"].get("pipeline_id", None)
if name is None:
name = os.path.basename[pipeline["source"]]
if "." in name:
name, _ = name.rsplit(".", maxsplit=1)
pipelines[name] = pipeline
return pipelines
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
op_configs = []
for pipeline_name, pipeline_data in self.pipeline_data.items():
pipeline_config = dict(pipeline_data["data"])
pipeline_id = pipeline_config.pop("pipeline_id", None)
doc = pipeline_config.pop("documentation", None)
# pipeline_config = PipelineConfig.from_config(data=pipeline_config, kiara=self._kiara)
op_details = PipelineOperationConfig(
pipeline_id=pipeline_id, pipeline_config=pipeline_config, doc=doc
)
op_configs.append(op_details)
return op_configs
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[PipelineOperationDetails]:
if isinstance(module, PipelineModule):
op_details = PipelineOperationDetails.create_operation_details(
operation_id=module.config.pipeline_id,
pipeline_inputs_schema=module.inputs_schema,
pipeline_outputs_schema=module.outputs_schema,
)
return op_details
else:
return None
pipeline_data
property
readonly
¶check_matching_operation(self, module)
¶Check whether the provided module is a valid operation for this type.
Source code in kiara/operations/included_core_operations/pipeline.py
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[PipelineOperationDetails]:
if isinstance(module, PipelineModule):
op_details = PipelineOperationDetails.create_operation_details(
operation_id=module.config.pipeline_id,
pipeline_inputs_schema=module.inputs_schema,
pipeline_outputs_schema=module.outputs_schema,
)
return op_details
else:
return None
retrieve_included_operation_configs(self)
¶Source code in kiara/operations/included_core_operations/pipeline.py
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
op_configs = []
for pipeline_name, pipeline_data in self.pipeline_data.items():
pipeline_config = dict(pipeline_data["data"])
pipeline_id = pipeline_config.pop("pipeline_id", None)
doc = pipeline_config.pop("documentation", None)
# pipeline_config = PipelineConfig.from_config(data=pipeline_config, kiara=self._kiara)
op_details = PipelineOperationConfig(
pipeline_id=pipeline_id, pipeline_config=pipeline_config, doc=doc
)
op_configs.append(op_details)
return op_configs
render_value
¶
RenderValueDetails (BaseOperationDetails)
pydantic-model
¶Source code in kiara/operations/included_core_operations/render_value.py
class RenderValueDetails(BaseOperationDetails):
source_type: str = Field(description="The type of the value to be rendered.")
target_type: str = Field(description="The type of the render result.")
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {
"value": {"type": "any", "doc": "The value to persist."},
"render_type": {
"type": "string",
"doc": "The render target/type of render output.",
},
"render_config": {
"type": "dict",
"doc": "A value type specific configuration for how to render the data.",
"optional": True,
},
}
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {"rendered_value": {"type": "any", "doc": "The rendered value."}}
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return {
self.source_type: inputs["value"],
"render_config": inputs.get("render_config", None),
}
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return {"rendered_value": outputs.get_value_obj("rendered_value")}
source_type: str
pydantic-field
required
¶The type of the value to be rendered.
target_type: str
pydantic-field
required
¶The type of the render result.
create_module_inputs(self, inputs)
¶Source code in kiara/operations/included_core_operations/render_value.py
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
return {
self.source_type: inputs["value"],
"render_config": inputs.get("render_config", None),
}
create_operation_outputs(self, outputs)
¶Source code in kiara/operations/included_core_operations/render_value.py
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
return {"rendered_value": outputs.get_value_obj("rendered_value")}
retrieve_inputs_schema(self)
¶Source code in kiara/operations/included_core_operations/render_value.py
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {
"value": {"type": "any", "doc": "The value to persist."},
"render_type": {
"type": "string",
"doc": "The render target/type of render output.",
},
"render_config": {
"type": "dict",
"doc": "A value type specific configuration for how to render the data.",
"optional": True,
},
}
retrieve_outputs_schema(self)
¶Source code in kiara/operations/included_core_operations/render_value.py
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {"rendered_value": {"type": "any", "doc": "The rendered value."}}
RenderValueOperationType (OperationType)
¶An operation that takes a value, and renders into a format that can be printed for output..
For a module profile to be picked up by this operation type, it needs to have: - exactly one output field named "rendered_value" - exactly two input fields, one of them named after the type it supports, and the other called 'render_config', of type 'dict'
Source code in kiara/operations/included_core_operations/render_value.py
class RenderValueOperationType(OperationType[RenderValueDetails]):
"""An operation that takes a value, and renders into a format that can be printed for output..
For a module profile to be picked up by this operation type, it needs to have:
- exactly one output field named "rendered_value"
- exactly two input fields, one of them named after the type it supports, and the other called 'render_config', of type 'dict'
"""
_operation_type_name = "render_value"
def _calculate_op_id(self, source_type: str, target_type: str):
if source_type == "any":
operation_id = f"render.as.{target_type}"
else:
operation_id = f"render.{source_type}.as.{target_type}"
return operation_id
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = {}
for name, module_cls in self._kiara.module_type_classes.items():
if not issubclass(module_cls, RenderValueModule):
continue
for (
source_type,
target_type,
) in module_cls.retrieve_supported_render_combinations():
if source_type not in self._kiara.data_type_names:
log_message("ignore.operation_config", operation_type="render_value", module_type=module_cls._module_type_name, source_type=source_type, target_type=target_type, reason=f"Source type '{source_type}' not registered.") # type: ignore
continue
if target_type not in self._kiara.data_type_names:
log_message(
"ignore.operation_config",
operation_type="render_value",
module_type=module_cls._module_type_name,
source_type=source_type, # type: ignore
target_type=target_type,
reason=f"Target type '{target_type}' not registered.",
)
continue
func_name = f"render__{source_type}__as__{target_type}"
attr = getattr(module_cls, func_name)
doc = DocumentationMetadataModel.from_function(attr)
mc = {"source_type": source_type, "target_type": target_type}
oc = ManifestOperationConfig(
module_type=name, module_config=mc, doc=doc
)
op_id = self._calculate_op_id(
source_type=source_type, target_type=target_type
)
result[op_id] = oc
for data_type_name, data_type_class in self._kiara.data_type_classes.items():
for attr in dir(data_type_class):
if not attr.startswith("render_as__"):
continue
target_type = attr[11:]
if target_type not in self._kiara.data_type_names:
log_message(
"operation_config.ignore",
operation_type="render_value",
module_type="value.extract_metadata",
source_type=data_type_name,
target_type=target_type,
reason=f"Target type '{target_type}' not registered.",
) # type: ignore
# TODO: inspect signature?
doc = DocumentationMetadataModel.from_string(
f"Render a {data_type_name} value as {target_type}."
)
mc = {
"source_type": data_type_name,
"target_type": target_type,
}
oc = ManifestOperationConfig(
module_type="value.render", module_config=mc, doc=doc
)
result[f"_type_{data_type_name}"] = oc
return result.values()
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[RenderValueDetails]:
details = self.extract_details(module)
if details is None:
return None
else:
return details
def extract_details(self, module: "KiaraModule") -> Optional[RenderValueDetails]:
if len(module.inputs_schema) != 2 or len(module.outputs_schema) != 1:
return None
target_type = None
for field_name, schema in module.outputs_schema.items():
if field_name != "rendered_value":
return None
target_type = schema.type
if target_type is None:
raise Exception("No target type available.")
input_field_match = None
render_config_match = None
for field_name, schema in module.inputs_schema.items():
if field_name == schema.type:
if input_field_match is not None:
# we can't deal (yet) with multiple fields
log_message(
"operation.ignore",
module=module.module_type_name,
reason=f"more than one input fields of type '{schema.type}'",
)
input_field_match = None
break
else:
input_field_match = field_name
elif field_name == "render_config":
render_config_match = field_name
if input_field_match is None:
return None
if render_config_match is None:
return None
input_field_type = module.inputs_schema[input_field_match].type
operation_id = self._calculate_op_id(
source_type=input_field_type, target_type=target_type
)
details = {
"operation_id": operation_id,
"source_type": input_field_type,
"target_type": target_type,
"is_internal_operation": True,
}
result = RenderValueDetails.create_operation_details(**details)
return result
def get_target_types_for(self, source_type: str) -> Mapping[str, Operation]:
# TODO: support for sub-types
result: Dict[str, Operation] = {}
for operation in self.operations.values():
details = self.retrieve_operation_details(operation)
if details.source_type == source_type:
target_type = details.target_type
if target_type in result.keys():
raise Exception(
f"More than one operation for render combination '{source_type}'/'{target_type}', this is not supported (for now)."
)
result[target_type] = operation
return result
def get_operation_for_render_combination(
self, source_type: str, target_type: str
) -> Operation:
type_lineage = self._kiara.type_registry.get_type_lineage(
data_type_name=source_type
)
for st in type_lineage:
target_types = self.get_target_types_for(source_type=st)
if not target_types:
continue
if target_type not in target_types.keys():
raise Exception(
f"No operation that produces '{target_type}' for source type: {st}."
)
return target_types[target_type]
raise Exception(f"No render opration(s) for source type: {source_type}.")
check_matching_operation(self, module)
¶Check whether the provided module is a valid operation for this type.
Source code in kiara/operations/included_core_operations/render_value.py
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[RenderValueDetails]:
details = self.extract_details(module)
if details is None:
return None
else:
return details
extract_details(self, module)
¶Source code in kiara/operations/included_core_operations/render_value.py
def extract_details(self, module: "KiaraModule") -> Optional[RenderValueDetails]:
if len(module.inputs_schema) != 2 or len(module.outputs_schema) != 1:
return None
target_type = None
for field_name, schema in module.outputs_schema.items():
if field_name != "rendered_value":
return None
target_type = schema.type
if target_type is None:
raise Exception("No target type available.")
input_field_match = None
render_config_match = None
for field_name, schema in module.inputs_schema.items():
if field_name == schema.type:
if input_field_match is not None:
# we can't deal (yet) with multiple fields
log_message(
"operation.ignore",
module=module.module_type_name,
reason=f"more than one input fields of type '{schema.type}'",
)
input_field_match = None
break
else:
input_field_match = field_name
elif field_name == "render_config":
render_config_match = field_name
if input_field_match is None:
return None
if render_config_match is None:
return None
input_field_type = module.inputs_schema[input_field_match].type
operation_id = self._calculate_op_id(
source_type=input_field_type, target_type=target_type
)
details = {
"operation_id": operation_id,
"source_type": input_field_type,
"target_type": target_type,
"is_internal_operation": True,
}
result = RenderValueDetails.create_operation_details(**details)
return result
get_operation_for_render_combination(self, source_type, target_type)
¶Source code in kiara/operations/included_core_operations/render_value.py
def get_operation_for_render_combination(
self, source_type: str, target_type: str
) -> Operation:
type_lineage = self._kiara.type_registry.get_type_lineage(
data_type_name=source_type
)
for st in type_lineage:
target_types = self.get_target_types_for(source_type=st)
if not target_types:
continue
if target_type not in target_types.keys():
raise Exception(
f"No operation that produces '{target_type}' for source type: {st}."
)
return target_types[target_type]
raise Exception(f"No render opration(s) for source type: {source_type}.")
get_target_types_for(self, source_type)
¶Source code in kiara/operations/included_core_operations/render_value.py
def get_target_types_for(self, source_type: str) -> Mapping[str, Operation]:
# TODO: support for sub-types
result: Dict[str, Operation] = {}
for operation in self.operations.values():
details = self.retrieve_operation_details(operation)
if details.source_type == source_type:
target_type = details.target_type
if target_type in result.keys():
raise Exception(
f"More than one operation for render combination '{source_type}'/'{target_type}', this is not supported (for now)."
)
result[target_type] = operation
return result
retrieve_included_operation_configs(self)
¶Source code in kiara/operations/included_core_operations/render_value.py
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = {}
for name, module_cls in self._kiara.module_type_classes.items():
if not issubclass(module_cls, RenderValueModule):
continue
for (
source_type,
target_type,
) in module_cls.retrieve_supported_render_combinations():
if source_type not in self._kiara.data_type_names:
log_message("ignore.operation_config", operation_type="render_value", module_type=module_cls._module_type_name, source_type=source_type, target_type=target_type, reason=f"Source type '{source_type}' not registered.") # type: ignore
continue
if target_type not in self._kiara.data_type_names:
log_message(
"ignore.operation_config",
operation_type="render_value",
module_type=module_cls._module_type_name,
source_type=source_type, # type: ignore
target_type=target_type,
reason=f"Target type '{target_type}' not registered.",
)
continue
func_name = f"render__{source_type}__as__{target_type}"
attr = getattr(module_cls, func_name)
doc = DocumentationMetadataModel.from_function(attr)
mc = {"source_type": source_type, "target_type": target_type}
oc = ManifestOperationConfig(
module_type=name, module_config=mc, doc=doc
)
op_id = self._calculate_op_id(
source_type=source_type, target_type=target_type
)
result[op_id] = oc
for data_type_name, data_type_class in self._kiara.data_type_classes.items():
for attr in dir(data_type_class):
if not attr.startswith("render_as__"):
continue
target_type = attr[11:]
if target_type not in self._kiara.data_type_names:
log_message(
"operation_config.ignore",
operation_type="render_value",
module_type="value.extract_metadata",
source_type=data_type_name,
target_type=target_type,
reason=f"Target type '{target_type}' not registered.",
) # type: ignore
# TODO: inspect signature?
doc = DocumentationMetadataModel.from_string(
f"Render a {data_type_name} value as {target_type}."
)
mc = {
"source_type": data_type_name,
"target_type": target_type,
}
oc = ManifestOperationConfig(
module_type="value.render", module_config=mc, doc=doc
)
result[f"_type_{data_type_name}"] = oc
return result.values()
serialize
¶
SerializeDetails (BaseOperationDetails)
pydantic-model
¶Source code in kiara/operations/included_core_operations/serialize.py
class SerializeDetails(BaseOperationDetails):
value_input_field: str = Field(
description="The (input) field name containing the value to be serialized."
)
value_input_type: str = Field(description="The type of the value to be serialized.")
serialized_value_output_field: str = Field(
description="The (output) field name containing the serialzied form of the value."
)
serialization_format: str = Field(
description="The name of the serialization format."
)
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {"value": {"type": "any", "doc": "The value to serialzie."}}
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {
"serialized_value": {
"type": "serialized_value",
"doc": "The serialized value details (and data).",
}
}
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
raise NotImplementedError()
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
raise NotImplementedError()
serialization_format: str
pydantic-field
required
¶The name of the serialization format.
serialized_value_output_field: str
pydantic-field
required
¶The (output) field name containing the serialzied form of the value.
value_input_field: str
pydantic-field
required
¶The (input) field name containing the value to be serialized.
value_input_type: str
pydantic-field
required
¶The type of the value to be serialized.
create_module_inputs(self, inputs)
¶Source code in kiara/operations/included_core_operations/serialize.py
def create_module_inputs(self, inputs: Mapping[str, Any]) -> Mapping[str, Any]:
raise NotImplementedError()
create_operation_outputs(self, outputs)
¶Source code in kiara/operations/included_core_operations/serialize.py
def create_operation_outputs(self, outputs: ValueMap) -> Mapping[str, Value]:
raise NotImplementedError()
retrieve_inputs_schema(self)
¶Source code in kiara/operations/included_core_operations/serialize.py
def retrieve_inputs_schema(self) -> ValueSetSchema:
return {"value": {"type": "any", "doc": "The value to serialzie."}}
retrieve_outputs_schema(self)
¶Source code in kiara/operations/included_core_operations/serialize.py
def retrieve_outputs_schema(self) -> ValueSetSchema:
return {
"serialized_value": {
"type": "serialized_value",
"doc": "The serialized value details (and data).",
}
}
SerializeOperationType (OperationType)
¶An operation that takes a value, and serializes it into the format suitable to the [serialized_value][kiara.data_types.included_core_types.SeriailzedValue] value type.
For a module profile to be picked up by this operation type, it needs to have:
- exactly one output field of type serialized_value
- either one of (in this order):
- exactly one input field
- one input field where the field name equals the type name
- an input field called 'value'
Source code in kiara/operations/included_core_operations/serialize.py
class SerializeOperationType(OperationType[SerializeDetails]):
"""An operation that takes a value, and serializes it into the format suitable to the [`serialized_value`][kiara.data_types.included_core_types.SeriailzedValue] value type.
For a module profile to be picked up by this operation type, it needs to have:
- exactly one output field of type `serialized_value`
- either one of (in this order):
- exactly one input field
- one input field where the field name equals the type name
- an input field called 'value'
"""
_operation_type_name = "serialize"
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = []
for name, module_cls in self._kiara.module_type_classes.items():
if not issubclass(module_cls, SerializeValueModule):
continue
for st in module_cls.retrieve_supported_source_types():
func_name = f"from__{st}"
attr = getattr(module_cls, func_name)
doc = DocumentationMetadataModel.from_function(attr)
mc = {"source_type": st}
oc = ManifestOperationConfig(
module_type=name, module_config=mc, doc=doc
)
result.append(oc)
return result
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[SerializeDetails]:
details = self.extract_details(module)
if details is None:
return None
else:
return details
def extract_details(self, module: KiaraModule) -> Optional[SerializeDetails]:
match = None
for field_name, schema in module.outputs_schema.items():
if schema.type != SERIALIZED_DATA_TYPE_NAME:
continue
else:
if match is not None:
log_message(
"ignore.operation",
reason=f"More than one field of type '{SERIALIZED_DATA_TYPE_NAME}'",
module_type=module.module_type_name,
)
continue
else:
match = field_name
if not match:
return None
if len(module.inputs_schema) == 1:
input_field: Optional[str] = next(iter(module.inputs_schema.keys()))
else:
input_field_match = None
for field_name, schema in module.inputs_schema.items():
if field_name == schema.type:
if input_field_match is not None:
input_field_match = None
break
else:
input_field_match = field_name
if input_field_match is not None:
input_field = input_field_match
elif "value" in module.inputs_schema.keys():
input_field = "value"
else:
input_field = None
if input_field is None:
return None
input_field_type = module.inputs_schema[input_field].type
value_schema: ValueSchema = module.outputs_schema[match]
serialized_value_type: SerializedValueType = self._kiara.type_registry.retrieve_data_type( # type: ignore # type: ignore
data_type_name=value_schema.type,
data_type_config=value_schema.type_config,
) # type: ignore
if input_field_type == "any":
operation_id = f"serialize.as.{serialized_value_type.format_name}"
else:
operation_id = (
f"serialize.{input_field_type}.as.{serialized_value_type.format_name}"
)
details: Dict[str, Any] = {
"operation_id": operation_id,
"value_input_field": input_field,
"value_input_type": input_field_type,
"serialized_value_output_field": match,
"serialization_format": serialized_value_type.format_name,
"is_internal_operation": True,
}
result = SerializeDetails.construct(**details)
return result
def find_serialzation_operation_for_type(self, type_name: str) -> Operation:
lineage = self._kiara.type_registry.get_type_lineage(type_name)
serialize_op: Optional[Operation] = None
for data_type in lineage:
match = []
op = None
for op in self.operations.values():
details = self.retrieve_operation_details(op)
if details.value_input_type == data_type:
match.append(op)
if match:
if len(match) > 1:
assert op is not None
raise Exception(
f"Multiple serialization operations found for type of '{op.operation_id}'. This is not supported (yet)."
)
serialize_op = match[0]
if serialize_op is None:
raise Exception(
f"Can't find serialization operation for type '{type_name}'."
)
return serialize_op
# def find_operation(self, **op_args: Any) -> Operation:
# op_conf = SerializeValueInputs(**op_args)
# input_value_type = op_conf.value.data_type_name
# op = self.find_serialzation_operation_for_type(input_value_type)
# return op
# def apply(self, inputs: SerializeValueInputs) -> SerializeValueOutputs:
#
# input_value_type = inputs.value.data_type_name
#
# op = self.find_serialzation_operation_for_type(input_value_type)
# op_details = self.retrieve_operation_details(op)
# op_inputs = {
# op_details.value_input_field: inputs.value
# }
#
# result = self._kiara.execute(manifest=op, inputs=op_inputs)
# op_details = self.retrieve_operation_details(op)
# result_data = {"serialized_value": result.get_value_obj(op_details.serialized_value_output_field)}
# return SerializeValueOutputs.construct(**result_data)
check_matching_operation(self, module)
¶Check whether the provided module is a valid operation for this type.
Source code in kiara/operations/included_core_operations/serialize.py
def check_matching_operation(
self, module: "KiaraModule"
) -> Optional[SerializeDetails]:
details = self.extract_details(module)
if details is None:
return None
else:
return details
extract_details(self, module)
¶Source code in kiara/operations/included_core_operations/serialize.py
def extract_details(self, module: KiaraModule) -> Optional[SerializeDetails]:
match = None
for field_name, schema in module.outputs_schema.items():
if schema.type != SERIALIZED_DATA_TYPE_NAME:
continue
else:
if match is not None:
log_message(
"ignore.operation",
reason=f"More than one field of type '{SERIALIZED_DATA_TYPE_NAME}'",
module_type=module.module_type_name,
)
continue
else:
match = field_name
if not match:
return None
if len(module.inputs_schema) == 1:
input_field: Optional[str] = next(iter(module.inputs_schema.keys()))
else:
input_field_match = None
for field_name, schema in module.inputs_schema.items():
if field_name == schema.type:
if input_field_match is not None:
input_field_match = None
break
else:
input_field_match = field_name
if input_field_match is not None:
input_field = input_field_match
elif "value" in module.inputs_schema.keys():
input_field = "value"
else:
input_field = None
if input_field is None:
return None
input_field_type = module.inputs_schema[input_field].type
value_schema: ValueSchema = module.outputs_schema[match]
serialized_value_type: SerializedValueType = self._kiara.type_registry.retrieve_data_type( # type: ignore # type: ignore
data_type_name=value_schema.type,
data_type_config=value_schema.type_config,
) # type: ignore
if input_field_type == "any":
operation_id = f"serialize.as.{serialized_value_type.format_name}"
else:
operation_id = (
f"serialize.{input_field_type}.as.{serialized_value_type.format_name}"
)
details: Dict[str, Any] = {
"operation_id": operation_id,
"value_input_field": input_field,
"value_input_type": input_field_type,
"serialized_value_output_field": match,
"serialization_format": serialized_value_type.format_name,
"is_internal_operation": True,
}
result = SerializeDetails.construct(**details)
return result
find_serialzation_operation_for_type(self, type_name)
¶Source code in kiara/operations/included_core_operations/serialize.py
def find_serialzation_operation_for_type(self, type_name: str) -> Operation:
lineage = self._kiara.type_registry.get_type_lineage(type_name)
serialize_op: Optional[Operation] = None
for data_type in lineage:
match = []
op = None
for op in self.operations.values():
details = self.retrieve_operation_details(op)
if details.value_input_type == data_type:
match.append(op)
if match:
if len(match) > 1:
assert op is not None
raise Exception(
f"Multiple serialization operations found for type of '{op.operation_id}'. This is not supported (yet)."
)
serialize_op = match[0]
if serialize_op is None:
raise Exception(
f"Can't find serialization operation for type '{type_name}'."
)
return serialize_op
retrieve_included_operation_configs(self)
¶Source code in kiara/operations/included_core_operations/serialize.py
def retrieve_included_operation_configs(
self,
) -> Iterable[Union[Mapping, OperationConfig]]:
result = []
for name, module_cls in self._kiara.module_type_classes.items():
if not issubclass(module_cls, SerializeValueModule):
continue
for st in module_cls.retrieve_supported_source_types():
func_name = f"from__{st}"
attr = getattr(module_cls, func_name)
doc = DocumentationMetadataModel.from_function(attr)
mc = {"source_type": st}
oc = ManifestOperationConfig(
module_type=name, module_config=mc, doc=doc
)
result.append(oc)
return result
processing
special
¶
Classes¶
JobStatusListener (Protocol)
¶
Source code in kiara/processing/__init__.py
class JobStatusListener(Protocol):
def job_status_changed(
self, job_id: uuid.UUID, old_status: Optional[JobStatus], new_status: JobStatus
):
pass
job_status_changed(self, job_id, old_status, new_status)
¶Source code in kiara/processing/__init__.py
def job_status_changed(
self, job_id: uuid.UUID, old_status: Optional[JobStatus], new_status: JobStatus
):
pass
ModuleProcessor (ABC)
¶
Source code in kiara/processing/__init__.py
class ModuleProcessor(abc.ABC):
def __init__(self, kiara: "Kiara"):
self._kiara: Kiara = kiara
self._created_jobs: Dict[uuid.UUID, Dict[str, Any]] = {}
self._active_jobs: Dict[uuid.UUID, ActiveJob] = {}
self._failed_jobs: Dict[uuid.UUID, ActiveJob] = {}
self._finished_jobs: Dict[uuid.UUID, ActiveJob] = {}
self._output_refs: Dict[uuid.UUID, ValueMapWritable] = {}
self._job_records: Dict[uuid.UUID, JobRecord] = {}
self._listeners: List[JobStatusListener] = []
def _send_job_event(
self, job_id: uuid.UUID, old_status: Optional[JobStatus], new_status: JobStatus
):
for listener in self._listeners:
listener.job_status_changed(
job_id=job_id, old_status=old_status, new_status=new_status
)
def register_job_status_listener(self, listener: JobStatusListener):
self._listeners.append(listener)
def get_job(self, job_id: uuid.UUID) -> ActiveJob:
if job_id in self._active_jobs.keys():
return self._active_jobs[job_id]
elif job_id in self._finished_jobs.keys():
return self._finished_jobs[job_id]
elif job_id in self._failed_jobs.keys():
return self._failed_jobs[job_id]
else:
raise Exception(f"No job with id '{job_id}' registered.")
def get_job_status(self, job_id: uuid.UUID) -> JobStatus:
job = self.get_job(job_id=job_id)
return job.status
def get_job_record(self, job_id: uuid.UUID) -> JobRecord:
if job_id in self._job_records.keys():
return self._job_records[job_id]
else:
raise Exception(f"No job record for job with id '{job_id}' registered.")
def create_job(self, job_config: JobConfig) -> uuid.UUID:
environments = {
env_name: env.model_data_hash
for env_name, env in self._kiara.current_environments.items()
}
result_pedigree = ValuePedigree(
kiara_id=self._kiara.id,
module_type=job_config.module_type,
module_config=job_config.module_config,
inputs=job_config.inputs,
environments=environments,
)
module = self._kiara.create_module(manifest=job_config)
outputs = ValueMapWritable.create_from_schema(
kiara=self._kiara, schema=module.outputs_schema, pedigree=result_pedigree
)
job_id = ID_REGISTRY.generate(kiara_id=self._kiara.id)
job_log = JobLog()
job = ActiveJob.construct(job_id=job_id, job_config=job_config, job_log=job_log)
ID_REGISTRY.update_metadata(job_id, obj=job)
job.job_log.add_log("job created")
job_details = {
"job_config": job_config,
"job": job,
"module": module,
"outputs": outputs,
}
self._created_jobs[job_id] = job_details
self._send_job_event(
job_id=job_id, old_status=None, new_status=JobStatus.CREATED
)
return job_id
def queue_job(self, job_id: uuid.UUID) -> ActiveJob:
job_details = self._created_jobs.pop(job_id)
job_config = job_details.pop("job_config")
job = job_details.pop("job")
module = job_details.pop("module")
outputs = job_details.pop("outputs")
self._active_jobs[job_id] = job
self._output_refs[job_id] = outputs
input_values = self._kiara.data_registry.load_values(job_config.inputs)
if module.is_pipeline():
module._set_job_registry(self._kiara.job_registry)
try:
self._add_processing_task(
job_id=job_id,
module=module,
inputs=input_values,
outputs=outputs,
job_log=job.job_log,
)
return job
except Exception as e:
job.error = str(e)
if is_debug():
try:
import traceback
traceback.print_exc()
except Exception:
pass
if isinstance(e, KiaraProcessingException):
e._module = module
e._inputs = ValueMapReadOnly.create_from_ids(
data_registry=self._kiara.data_registry, **job_config.inputs
)
job._exception = e
raise e
else:
kpe = KiaraProcessingException(
e,
module=module,
inputs=ValueMapReadOnly.create_from_ids(
self._kiara.data_registry, **job_config.inputs
),
)
job._exception = kpe
raise kpe
def job_status_updated(
self, job_id: uuid.UUID, status: Union[JobStatus, str, Exception]
):
job = self._active_jobs.get(job_id, None)
if job is None:
raise Exception(
f"Can't retrieve active job with id '{job_id}', no such job registered."
)
old_status = job.status
if status == JobStatus.SUCCESS:
self._active_jobs.pop(job_id)
job.job_log.add_log("job finished successfully")
job.status = JobStatus.SUCCESS
job.finished = datetime.now()
values = self._output_refs[job_id]
values.sync_values()
value_ids = values.get_all_value_ids()
job.results = value_ids
job.job_log.percent_finished = 100
job_record = JobRecord.from_active_job(active_job=job)
self._job_records[job_id] = job_record
self._finished_jobs[job_id] = job
elif status == JobStatus.FAILED or isinstance(status, (str, Exception)):
self._active_jobs.pop(job_id)
job.job_log.add_log("job failed")
job.status = JobStatus.FAILED
job.finished = datetime.now()
if isinstance(status, str):
job.error = status
elif isinstance(status, Exception):
job.error = str(status)
job._exception = status
self._failed_jobs[job_id] = job
elif status == JobStatus.STARTED:
job.job_log.add_log("job started")
job.status = JobStatus.STARTED
job.started = datetime.now()
else:
raise ValueError(f"Invalid value for status: {status}")
self._send_job_event(
job_id=job_id, old_status=old_status, new_status=job.status
)
def wait_for(self, *job_ids: uuid.UUID):
"""Wait for the jobs with the specified ids, also optionally sync their outputs with the pipeline value state."""
self._wait_for(*job_ids)
for job_id in job_ids:
job = self._job_records.get(job_id, None)
if job is None:
_job = self._failed_jobs.get(job_id, None)
if _job is None:
raise Exception(f"Can't find job with id: {job_id}")
@abc.abstractmethod
def _wait_for(self, *job_ids: uuid.UUID):
pass
@abc.abstractmethod
def _add_processing_task(
self,
job_id: uuid.UUID,
module: "KiaraModule",
inputs: ValueMap,
outputs: ValueMapWritable,
job_log: JobLog,
) -> str:
pass
Methods¶
create_job(self, job_config)
¶Source code in kiara/processing/__init__.py
def create_job(self, job_config: JobConfig) -> uuid.UUID:
environments = {
env_name: env.model_data_hash
for env_name, env in self._kiara.current_environments.items()
}
result_pedigree = ValuePedigree(
kiara_id=self._kiara.id,
module_type=job_config.module_type,
module_config=job_config.module_config,
inputs=job_config.inputs,
environments=environments,
)
module = self._kiara.create_module(manifest=job_config)
outputs = ValueMapWritable.create_from_schema(
kiara=self._kiara, schema=module.outputs_schema, pedigree=result_pedigree
)
job_id = ID_REGISTRY.generate(kiara_id=self._kiara.id)
job_log = JobLog()
job = ActiveJob.construct(job_id=job_id, job_config=job_config, job_log=job_log)
ID_REGISTRY.update_metadata(job_id, obj=job)
job.job_log.add_log("job created")
job_details = {
"job_config": job_config,
"job": job,
"module": module,
"outputs": outputs,
}
self._created_jobs[job_id] = job_details
self._send_job_event(
job_id=job_id, old_status=None, new_status=JobStatus.CREATED
)
return job_id
get_job(self, job_id)
¶Source code in kiara/processing/__init__.py
def get_job(self, job_id: uuid.UUID) -> ActiveJob:
if job_id in self._active_jobs.keys():
return self._active_jobs[job_id]
elif job_id in self._finished_jobs.keys():
return self._finished_jobs[job_id]
elif job_id in self._failed_jobs.keys():
return self._failed_jobs[job_id]
else:
raise Exception(f"No job with id '{job_id}' registered.")
get_job_record(self, job_id)
¶Source code in kiara/processing/__init__.py
def get_job_record(self, job_id: uuid.UUID) -> JobRecord:
if job_id in self._job_records.keys():
return self._job_records[job_id]
else:
raise Exception(f"No job record for job with id '{job_id}' registered.")
get_job_status(self, job_id)
¶Source code in kiara/processing/__init__.py
def get_job_status(self, job_id: uuid.UUID) -> JobStatus:
job = self.get_job(job_id=job_id)
return job.status
job_status_updated(self, job_id, status)
¶Source code in kiara/processing/__init__.py
def job_status_updated(
self, job_id: uuid.UUID, status: Union[JobStatus, str, Exception]
):
job = self._active_jobs.get(job_id, None)
if job is None:
raise Exception(
f"Can't retrieve active job with id '{job_id}', no such job registered."
)
old_status = job.status
if status == JobStatus.SUCCESS:
self._active_jobs.pop(job_id)
job.job_log.add_log("job finished successfully")
job.status = JobStatus.SUCCESS
job.finished = datetime.now()
values = self._output_refs[job_id]
values.sync_values()
value_ids = values.get_all_value_ids()
job.results = value_ids
job.job_log.percent_finished = 100
job_record = JobRecord.from_active_job(active_job=job)
self._job_records[job_id] = job_record
self._finished_jobs[job_id] = job
elif status == JobStatus.FAILED or isinstance(status, (str, Exception)):
self._active_jobs.pop(job_id)
job.job_log.add_log("job failed")
job.status = JobStatus.FAILED
job.finished = datetime.now()
if isinstance(status, str):
job.error = status
elif isinstance(status, Exception):
job.error = str(status)
job._exception = status
self._failed_jobs[job_id] = job
elif status == JobStatus.STARTED:
job.job_log.add_log("job started")
job.status = JobStatus.STARTED
job.started = datetime.now()
else:
raise ValueError(f"Invalid value for status: {status}")
self._send_job_event(
job_id=job_id, old_status=old_status, new_status=job.status
)
queue_job(self, job_id)
¶Source code in kiara/processing/__init__.py
def queue_job(self, job_id: uuid.UUID) -> ActiveJob:
job_details = self._created_jobs.pop(job_id)
job_config = job_details.pop("job_config")
job = job_details.pop("job")
module = job_details.pop("module")
outputs = job_details.pop("outputs")
self._active_jobs[job_id] = job
self._output_refs[job_id] = outputs
input_values = self._kiara.data_registry.load_values(job_config.inputs)
if module.is_pipeline():
module._set_job_registry(self._kiara.job_registry)
try:
self._add_processing_task(
job_id=job_id,
module=module,
inputs=input_values,
outputs=outputs,
job_log=job.job_log,
)
return job
except Exception as e:
job.error = str(e)
if is_debug():
try:
import traceback
traceback.print_exc()
except Exception:
pass
if isinstance(e, KiaraProcessingException):
e._module = module
e._inputs = ValueMapReadOnly.create_from_ids(
data_registry=self._kiara.data_registry, **job_config.inputs
)
job._exception = e
raise e
else:
kpe = KiaraProcessingException(
e,
module=module,
inputs=ValueMapReadOnly.create_from_ids(
self._kiara.data_registry, **job_config.inputs
),
)
job._exception = kpe
raise kpe
register_job_status_listener(self, listener)
¶Source code in kiara/processing/__init__.py
def register_job_status_listener(self, listener: JobStatusListener):
self._listeners.append(listener)
wait_for(self, *job_ids)
¶Wait for the jobs with the specified ids, also optionally sync their outputs with the pipeline value state.
Source code in kiara/processing/__init__.py
def wait_for(self, *job_ids: uuid.UUID):
"""Wait for the jobs with the specified ids, also optionally sync their outputs with the pipeline value state."""
self._wait_for(*job_ids)
for job_id in job_ids:
job = self._job_records.get(job_id, None)
if job is None:
_job = self._failed_jobs.get(job_id, None)
if _job is None:
raise Exception(f"Can't find job with id: {job_id}")
ProcessorConfig (BaseModel)
pydantic-model
¶
Source code in kiara/processing/__init__.py
class ProcessorConfig(BaseModel):
module_processor_type: Literal["synchronous", "multi-threaded"] = "synchronous"
module_processor_type: Literal['synchronous', 'multi-threaded']
pydantic-field
¶
synchronous
¶
SynchronousProcessor (ModuleProcessor)
¶Source code in kiara/processing/synchronous.py
class SynchronousProcessor(ModuleProcessor):
def _add_processing_task(
self,
job_id: uuid.UUID,
module: "KiaraModule",
inputs: ValueMap,
outputs: ValueMapWritable,
job_log: JobLog,
):
self.job_status_updated(job_id=job_id, status=JobStatus.STARTED)
try:
module.process_step(inputs=inputs, outputs=outputs, job_log=job_log)
# output_wrap._sync()
self.job_status_updated(job_id=job_id, status=JobStatus.SUCCESS)
except Exception as e:
self.job_status_updated(job_id=job_id, status=e)
def _wait_for(self, *job_ids: uuid.UUID):
# jobs will always be finished, since we were waiting for them in the 'process' method
return
SynchronousProcessorConfig (ProcessorConfig)
pydantic-model
¶Source code in kiara/processing/synchronous.py
class SynchronousProcessorConfig(ProcessorConfig):
pass
registries
special
¶
ARCHIVE_CONFIG_CLS
¶
Classes¶
ArchiveConfig (BaseModel)
pydantic-model
¶
Source code in kiara/registries/__init__.py
class ArchiveConfig(BaseModel):
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
Config
¶Source code in kiara/registries/__init__.py
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
json_loads
¶json_dumps(v, *, default=None, **args)
¶Source code in kiara/registries/__init__.py
def orjson_dumps(v, *, default=None, **args):
# orjson.dumps returns bytes, to match standard json.dumps we need to decode
try:
return orjson.dumps(v, default=default, **args).decode()
except Exception as e:
if is_debug():
print(f"Error dumping json data: {e}")
from kiara import dbg
dbg(v)
raise e
BaseArchive (KiaraArchive, Generic)
¶
Source code in kiara/registries/__init__.py
class BaseArchive(KiaraArchive, Generic[ARCHIVE_CONFIG_CLS]):
_config_cls = ArchiveConfig
def __init__(self, archive_id: uuid.UUID, config: ARCHIVE_CONFIG_CLS):
self._archive_id: uuid.UUID = archive_id
self._config: ARCHIVE_CONFIG_CLS = config
self._kiara: Optional["Kiara"] = None
@property
def config(self) -> ARCHIVE_CONFIG_CLS:
return self._config
@property
def archive_id(self) -> uuid.UUID:
return self._archive_id
@property
def kiara_context(self) -> "Kiara":
if self._kiara is None:
raise Exception("Archive not registered into a kiara context yet.")
return self._kiara
def register_archive(self, kiara: "Kiara") -> uuid.UUID:
self._kiara = kiara
return self.archive_id
archive_id: UUID
property
readonly
¶config: ~ARCHIVE_CONFIG_CLS
property
readonly
¶kiara_context: Kiara
property
readonly
¶
_config_cls (BaseModel)
private
pydantic-model
¶Source code in kiara/registries/__init__.py
class ArchiveConfig(BaseModel):
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
Config
¶Source code in kiara/registries/__init__.py
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
json_loads
¶json_dumps(v, *, default=None, **args)
¶Source code in kiara/registries/__init__.py
def orjson_dumps(v, *, default=None, **args):
# orjson.dumps returns bytes, to match standard json.dumps we need to decode
try:
return orjson.dumps(v, default=default, **args).decode()
except Exception as e:
if is_debug():
print(f"Error dumping json data: {e}")
from kiara import dbg
dbg(v)
raise e
register_archive(self, kiara)
¶Source code in kiara/registries/__init__.py
def register_archive(self, kiara: "Kiara") -> uuid.UUID:
self._kiara = kiara
return self.archive_id
FileSystemArchiveConfig (ArchiveConfig)
pydantic-model
¶
KiaraArchive (ABC)
¶
Source code in kiara/registries/__init__.py
class KiaraArchive(abc.ABC):
_config_cls: ClassVar[Type[ArchiveConfig]] = ArchiveConfig
@classmethod
@abc.abstractmethod
def supported_item_types(cls) -> Iterable[str]:
pass
@classmethod
@abc.abstractmethod
def is_writeable(cls) -> bool:
pass
@abc.abstractmethod
def register_archive(self, kiara: "Kiara") -> uuid.UUID:
pass
_config_cls (BaseModel)
private
pydantic-model
¶Source code in kiara/registries/__init__.py
class ArchiveConfig(BaseModel):
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
Config
¶Source code in kiara/registries/__init__.py
class Config:
json_loads = orjson.loads
json_dumps = orjson_dumps
json_loads
¶json_dumps(v, *, default=None, **args)
¶Source code in kiara/registries/__init__.py
def orjson_dumps(v, *, default=None, **args):
# orjson.dumps returns bytes, to match standard json.dumps we need to decode
try:
return orjson.dumps(v, default=default, **args).decode()
except Exception as e:
if is_debug():
print(f"Error dumping json data: {e}")
from kiara import dbg
dbg(v)
raise e
is_writeable()
classmethod
¶Source code in kiara/registries/__init__.py
@classmethod
@abc.abstractmethod
def is_writeable(cls) -> bool:
pass
register_archive(self, kiara)
¶Source code in kiara/registries/__init__.py
@abc.abstractmethod
def register_archive(self, kiara: "Kiara") -> uuid.UUID:
pass
supported_item_types()
classmethod
¶Source code in kiara/registries/__init__.py
@classmethod
@abc.abstractmethod
def supported_item_types(cls) -> Iterable[str]:
pass
Modules¶
aliases
special
¶
logger
¶Classes¶
AliasArchive (BaseArchive)
¶Source code in kiara/registries/aliases/__init__.py
class AliasArchive(BaseArchive):
@classmethod
def supported_item_types(cls) -> Iterable[str]:
return ["alias"]
@abc.abstractmethod
def retrieve_all_aliases(self) -> Optional[Mapping[str, uuid.UUID]]:
"""Retrieve a list of all aliases registered in this archive.
The result of this method can be 'None', for cases where the aliases are determined dynamically.
In kiara, the result of this method is mostly used to improve performance when looking up an alias.
Returns:
a list of strings (the aliases), or 'None' if this archive does not support alias indexes.
"""
@abc.abstractmethod
def find_value_id_for_alias(self, alias: str) -> Optional[uuid.UUID]:
pass
@abc.abstractmethod
def find_aliases_for_value_id(self, value_id: uuid.UUID) -> Optional[Set[str]]:
pass
@classmethod
def is_writeable(cls) -> bool:
return False
find_aliases_for_value_id(self, value_id)
¶Source code in kiara/registries/aliases/__init__.py
@abc.abstractmethod
def find_aliases_for_value_id(self, value_id: uuid.UUID) -> Optional[Set[str]]:
pass
find_value_id_for_alias(self, alias)
¶Source code in kiara/registries/aliases/__init__.py
@abc.abstractmethod
def find_value_id_for_alias(self, alias: str) -> Optional[uuid.UUID]:
pass
is_writeable()
classmethod
¶Source code in kiara/registries/aliases/__init__.py
@classmethod
def is_writeable(cls) -> bool:
return False
retrieve_all_aliases(self)
¶Retrieve a list of all aliases registered in this archive.
The result of this method can be 'None', for cases where the aliases are determined dynamically. In kiara, the result of this method is mostly used to improve performance when looking up an alias.
Returns:
| Type | Description |
|---|---|
Optional[Mapping[str, uuid.UUID]] |
a list of strings (the aliases), or 'None' if this archive does not support alias indexes. |
Source code in kiara/registries/aliases/__init__.py
@abc.abstractmethod
def retrieve_all_aliases(self) -> Optional[Mapping[str, uuid.UUID]]:
"""Retrieve a list of all aliases registered in this archive.
The result of this method can be 'None', for cases where the aliases are determined dynamically.
In kiara, the result of this method is mostly used to improve performance when looking up an alias.
Returns:
a list of strings (the aliases), or 'None' if this archive does not support alias indexes.
"""
supported_item_types()
classmethod
¶Source code in kiara/registries/aliases/__init__.py
@classmethod
def supported_item_types(cls) -> Iterable[str]:
return ["alias"]
AliasItem (tuple)
¶AliasItem(full_alias, rel_alias, value_id, alias_archive, alias_archive_id)
Source code in kiara/registries/aliases/__init__.py
class AliasItem(NamedTuple):
full_alias: str
rel_alias: str
value_id: uuid.UUID
alias_archive: str
alias_archive_id: uuid.UUID
AliasRegistry
¶Source code in kiara/registries/aliases/__init__.py
class AliasRegistry(object):
def __init__(self, kiara: "Kiara"):
self._kiara: Kiara = kiara
self._event_callback: Callable = self._kiara.event_registry.add_producer(self)
self._alias_archives: Dict[str, AliasArchive] = {}
"""All registered archives/stores."""
self._default_alias_store: Optional[str] = None
"""The alias of the store where new aliases are stored by default."""
self._cached_aliases: Optional[Dict[str, AliasItem]] = None
self._cached_aliases_by_id: Optional[Dict[uuid.UUID, Set[AliasItem]]] = None
def register_archive(
self,
archive: AliasArchive,
alias: str = None,
set_as_default_store: Optional[bool] = None,
):
alias_archive_id = archive.register_archive(kiara=self._kiara)
if alias is None:
alias = str(alias_archive_id)
if "." in alias:
raise Exception(
f"Can't register alias archive with as '{alias}': registered name is not allowed to contain a '.' character (yet)."
)
if alias in self._alias_archives.keys():
raise Exception(f"Can't add store, alias '{alias}' already registered.")
self._alias_archives[alias] = archive
is_store = False
is_default_store = False
if isinstance(archive, AliasStore):
is_store = True
if set_as_default_store and self._default_alias_store is not None:
raise Exception(
f"Can't set alias store '{alias}' as default store: default store already set."
)
if self._default_alias_store is None:
is_default_store = True
self._default_alias_store = alias
event = AliasArchiveAddedEvent.construct(
kiara_id=self._kiara.id,
alias_archive_id=archive.archive_id,
alias_archive_alias=alias,
is_store=is_store,
is_default_store=is_default_store,
)
self._event_callback(event)
@property
def default_alias_store(self) -> str:
if self._default_alias_store is None:
raise Exception("No default alias store set (yet).")
return self._default_alias_store
def get_archive(self, archive_id: Optional[str] = None) -> AliasArchive:
if archive_id is None:
archive_id = self.default_alias_store
if archive_id is None:
raise Exception("Can't retrieve default alias archive, none set (yet).")
archive = self._alias_archives.get(archive_id, None)
if archive is None:
raise Exception(f"No archive registered for base alias '{archive_id}'.")
return archive
@property
def all_aliases(self) -> Iterable[str]:
return self.aliases.keys()
@property
def aliases_by_id(self) -> Mapping[uuid.UUID, Set[AliasItem]]:
if self._cached_aliases_by_id is None:
self.aliases # noqa
return self._cached_aliases_by_id # type: ignore
@property
def aliases(self) -> Dict[str, AliasItem]:
"""Retrieve a map of all available aliases, context wide, with the registered archive aliases as values."""
if self._cached_aliases is not None:
return self._cached_aliases
# TODO: multithreading lock
all_aliases: Dict[str, AliasItem] = {}
all_aliases_by_id: Dict[uuid.UUID, Set[AliasItem]] = {}
for archive_alias, archive in self._alias_archives.items():
alias_map = archive.retrieve_all_aliases()
if alias_map is None:
continue
for alias, v_id in alias_map.items():
if archive_alias == self.default_alias_store:
final_alias = alias
else:
final_alias = f"{archive_alias}.{alias}"
if final_alias in all_aliases.keys():
raise Exception(
f"Inconsistent alias registry: alias '{final_alias}' available more than once."
)
item = AliasItem(
full_alias=final_alias,
rel_alias=alias,
value_id=v_id,
alias_archive=archive_alias,
alias_archive_id=archive.archive_id,
)
all_aliases[final_alias] = item
all_aliases_by_id.setdefault(v_id, set()).add(item)
self._cached_aliases = all_aliases
self._cached_aliases_by_id = all_aliases_by_id
return self._cached_aliases
def find_value_id_for_alias(self, alias: str) -> Optional[uuid.UUID]:
alias_item = self.aliases.get(alias, None)
if alias_item is not None:
return alias_item.value_id
if "." not in alias:
return None
archive_id, rest = alias.split(".", maxsplit=2)
archive = self.get_archive(archive_id=archive_id)
v_id = archive.find_value_id_for_alias(alias=rest)
# TODO: cache this?
return v_id
def find_aliases_for_value_id(
self, value_id: uuid.UUID, search_dynamic_archives: bool = False
) -> Set[str]:
aliases = set([a.full_alias for a in self.aliases_by_id.get(value_id, [])])
if search_dynamic_archives:
for archive_alias, archive in self._alias_archives.items():
_aliases = archive.find_aliases_for_value_id(value_id=value_id)
# TODO: cache those results
if _aliases:
for a in _aliases:
aliases.add(f"{archive_alias}.{a}")
return aliases
def register_aliases(self, value_id: uuid.UUID, *aliases: str):
store_name = self.default_alias_store
store: AliasStore = self.get_archive(archive_id=store_name) # type: ignore
self.aliases # noqu
store.register_aliases(value_id, *aliases)
for alias in aliases:
alias_item = AliasItem(
full_alias=alias,
rel_alias=alias,
value_id=value_id,
alias_archive=store_name,
alias_archive_id=store.archive_id,
)
if alias in self.aliases.keys():
logger.info("alias.replace", alias=alias)
# raise NotImplementedError()
self.aliases[alias] = alias_item
self._cached_aliases_by_id.setdefault(value_id, set()).add(alias_item) # type: ignore
aliases: Dict[str, kiara.registries.aliases.AliasItem]
property
readonly
¶Retrieve a map of all available aliases, context wide, with the registered archive aliases as values.
aliases_by_id: Mapping[uuid.UUID, Set[kiara.registries.aliases.AliasItem]]
property
readonly
¶all_aliases: Iterable[str]
property
readonly
¶default_alias_store: str
property
readonly
¶find_aliases_for_value_id(self, value_id, search_dynamic_archives=False)
¶Source code in kiara/registries/aliases/__init__.py
def find_aliases_for_value_id(
self, value_id: uuid.UUID, search_dynamic_archives: bool = False
) -> Set[str]:
aliases = set([a.full_alias for a in self.aliases_by_id.get(value_id, [])])
if search_dynamic_archives:
for archive_alias, archive in self._alias_archives.items():
_aliases = archive.find_aliases_for_value_id(value_id=value_id)
# TODO: cache those results
if _aliases:
for a in _aliases:
aliases.add(f"{archive_alias}.{a}")
return aliases
find_value_id_for_alias(self, alias)
¶Source code in kiara/registries/aliases/__init__.py
def find_value_id_for_alias(self, alias: str) -> Optional[uuid.UUID]:
alias_item = self.aliases.get(alias, None)
if alias_item is not None:
return alias_item.value_id
if "." not in alias:
return None
archive_id, rest = alias.split(".", maxsplit=2)
archive = self.get_archive(archive_id=archive_id)
v_id = archive.find_value_id_for_alias(alias=rest)
# TODO: cache this?
return v_id
get_archive(self, archive_id=None)
¶Source code in kiara/registries/aliases/__init__.py
def get_archive(self, archive_id: Optional[str] = None) -> AliasArchive:
if archive_id is None:
archive_id = self.default_alias_store
if archive_id is None:
raise Exception("Can't retrieve default alias archive, none set (yet).")
archive = self._alias_archives.get(archive_id, None)
if archive is None:
raise Exception(f"No archive registered for base alias '{archive_id}'.")
return archive
register_aliases(self, value_id, *aliases)
¶Source code in kiara/registries/aliases/__init__.py
def register_aliases(self, value_id: uuid.UUID, *aliases: str):
store_name = self.default_alias_store
store: AliasStore = self.get_archive(archive_id=store_name) # type: ignore
self.aliases # noqu
store.register_aliases(value_id, *aliases)
for alias in aliases:
alias_item = AliasItem(
full_alias=alias,
rel_alias=alias,
value_id=value_id,
alias_archive=store_name,
alias_archive_id=store.archive_id,
)
if alias in self.aliases.keys():
logger.info("alias.replace", alias=alias)
# raise NotImplementedError()
self.aliases[alias] = alias_item
self._cached_aliases_by_id.setdefault(value_id, set()).add(alias_item) # type: ignore
register_archive(self, archive, alias=None, set_as_default_store=None)
¶Source code in kiara/registries/aliases/__init__.py
def register_archive(
self,
archive: AliasArchive,
alias: str = None,
set_as_default_store: Optional[bool] = None,
):
alias_archive_id = archive.register_archive(kiara=self._kiara)
if alias is None:
alias = str(alias_archive_id)
if "." in alias:
raise Exception(
f"Can't register alias archive with as '{alias}': registered name is not allowed to contain a '.' character (yet)."
)
if alias in self._alias_archives.keys():
raise Exception(f"Can't add store, alias '{alias}' already registered.")
self._alias_archives[alias] = archive
is_store = False
is_default_store = False
if isinstance(archive, AliasStore):
is_store = True
if set_as_default_store and self._default_alias_store is not None:
raise Exception(
f"Can't set alias store '{alias}' as default store: default store already set."
)
if self._default_alias_store is None:
is_default_store = True
self._default_alias_store = alias
event = AliasArchiveAddedEvent.construct(
kiara_id=self._kiara.id,
alias_archive_id=archive.archive_id,
alias_archive_alias=alias,
is_store=is_store,
is_default_store=is_default_store,
)
self._event_callback(event)
AliasStore (AliasArchive)
¶Source code in kiara/registries/aliases/__init__.py
class AliasStore(AliasArchive):
@abc.abstractmethod
def register_aliases(self, value_id: uuid.UUID, *aliases: str):
pass
register_aliases(self, value_id, *aliases)
¶Source code in kiara/registries/aliases/__init__.py
@abc.abstractmethod
def register_aliases(self, value_id: uuid.UUID, *aliases: str):
pass
Modules¶
archives
¶
FileSystemAliasArchive (AliasArchive)
¶Source code in kiara/registries/aliases/archives.py
class FileSystemAliasArchive(AliasArchive):
_archive_type_name = "filesystem_alias_archive"
_config_cls = FileSystemArchiveConfig
def __init__(self, archive_id: uuid.UUID, config: ARCHIVE_CONFIG_CLS):
super().__init__(archive_id=archive_id, config=config)
self._base_path: Optional[Path] = None
@property
def alias_store_path(self) -> Path:
if self._base_path is not None:
return self._base_path
self._base_path = Path(self.config.base_path) / str(self.archive_id)
self._base_path.mkdir(parents=True, exist_ok=True)
return self._base_path
@property
def aliases_path(self) -> Path:
return self.alias_store_path / "aliases"
@property
def value_id_path(self) -> Path:
return self.alias_store_path / "value_ids"
def _translate_alias(self, alias: str) -> Path:
if "." in alias:
tokens = alias.split(".")
alias_path = (
self.aliases_path.joinpath(*tokens[0:-1]) / f"{tokens[-1]}.alias"
)
else:
alias_path = self.aliases_path / f"{alias}.alias"
return alias_path
def _translate_alias_path(self, alias_path: Path) -> str:
relative = alias_path.relative_to(self.aliases_path).as_posix()[:-6]
if os.path.sep not in relative:
alias = relative
else:
alias = ".".join(relative.split(os.path.sep))
return alias
def _translate_value_id(self, value_id: uuid.UUID) -> Path:
tokens = str(value_id).split("-")
value_id_path = (
self.value_id_path.joinpath(*tokens[0:-1]) / f"{tokens[-1]}.value"
)
return value_id_path
def _translate_value_path(self, value_path: Path) -> uuid.UUID:
relative = value_path.relative_to(self.value_id_path).as_posix()[:-6]
value_id_str = "-".join(relative.split(os.path.sep))
return uuid.UUID(value_id_str)
def retrieve_all_aliases(self) -> Mapping[str, uuid.UUID]:
all_aliases = self.aliases_path.rglob("*.alias")
result = {}
for alias_path in all_aliases:
alias = self._translate_alias_path(alias_path=alias_path)
value_id = self._find_value_id_for_alias_path(alias_path=alias_path)
assert value_id is not None
result[alias] = value_id
return result
def find_value_id_for_alias(self, alias: str) -> Optional[uuid.UUID]:
alias_path = self._translate_alias(alias)
if not alias_path.exists():
return None
return self._find_value_id_for_alias_path(alias_path=alias_path)
def _find_value_id_for_alias_path(self, alias_path: Path) -> Optional[uuid.UUID]:
resolved = alias_path.resolve()
assert resolved.name.endswith(".value")
value_id = self._translate_value_path(value_path=resolved)
return value_id
def find_aliases_for_value_id(self, value_id: uuid.UUID) -> Optional[Set[str]]:
raise NotImplementedError()
alias_store_path: Path
property
readonly
¶aliases_path: Path
property
readonly
¶value_id_path: Path
property
readonly
¶
_config_cls (ArchiveConfig)
private
pydantic-model
¶find_aliases_for_value_id(self, value_id)
¶Source code in kiara/registries/aliases/archives.py
def find_aliases_for_value_id(self, value_id: uuid.UUID) -> Optional[Set[str]]:
raise NotImplementedError()
find_value_id_for_alias(self, alias)
¶Source code in kiara/registries/aliases/archives.py
def find_value_id_for_alias(self, alias: str) -> Optional[uuid.UUID]:
alias_path = self._translate_alias(alias)
if not alias_path.exists():
return None
return self._find_value_id_for_alias_path(alias_path=alias_path)
retrieve_all_aliases(self)
¶Retrieve a list of all aliases registered in this archive.
The result of this method can be 'None', for cases where the aliases are determined dynamically. In kiara, the result of this method is mostly used to improve performance when looking up an alias.
Returns:
| Type | Description |
|---|---|
Mapping[str, uuid.UUID] |
a list of strings (the aliases), or 'None' if this archive does not support alias indexes. |
Source code in kiara/registries/aliases/archives.py
def retrieve_all_aliases(self) -> Mapping[str, uuid.UUID]:
all_aliases = self.aliases_path.rglob("*.alias")
result = {}
for alias_path in all_aliases:
alias = self._translate_alias_path(alias_path=alias_path)
value_id = self._find_value_id_for_alias_path(alias_path=alias_path)
assert value_id is not None
result[alias] = value_id
return result
FileSystemAliasStore (FileSystemAliasArchive, AliasStore)
¶Source code in kiara/registries/aliases/archives.py
class FileSystemAliasStore(FileSystemAliasArchive, AliasStore):
_archive_type_name = "filesystem_alias_store"
@classmethod
def is_writeable(cls) -> bool:
return True
def register_aliases(self, value_id: uuid.UUID, *aliases: str):
value_path = self._translate_value_id(value_id=value_id)
value_path.parent.mkdir(parents=True, exist_ok=True)
value_path.touch()
for alias in aliases:
alias_path = self._translate_alias(alias)
alias_path.parent.mkdir(parents=True, exist_ok=True)
if alias_path.exists():
resolved = alias_path.resolve()
if resolved == value_path:
continue
alias_path.unlink()
alias_path.symlink_to(value_path)
is_writeable()
classmethod
¶Source code in kiara/registries/aliases/archives.py
@classmethod
def is_writeable(cls) -> bool:
return True
register_aliases(self, value_id, *aliases)
¶Source code in kiara/registries/aliases/archives.py
def register_aliases(self, value_id: uuid.UUID, *aliases: str):
value_path = self._translate_value_id(value_id=value_id)
value_path.parent.mkdir(parents=True, exist_ok=True)
value_path.touch()
for alias in aliases:
alias_path = self._translate_alias(alias)
alias_path.parent.mkdir(parents=True, exist_ok=True)
if alias_path.exists():
resolved = alias_path.resolve()
if resolved == value_path:
continue
alias_path.unlink()
alias_path.symlink_to(value_path)
data
special
¶
logger
¶Classes¶
DataRegistry
¶Source code in kiara/registries/data/__init__.py
class DataRegistry(object):
def __init__(self, kiara: "Kiara"):
self._kiara: Kiara = kiara
self._engine: Engine = self._kiara._engine
self._event_callback: Callable = self._kiara.event_registry.add_producer(self)
self._data_archives: Dict[str, DataArchive] = {}
self._default_data_store: Optional[str] = None
self._registered_values: Dict[uuid.UUID, Value] = {}
self._value_archive_lookup_map: Dict[uuid.UUID, str] = {}
self._values_by_hash: Dict[int, Set[uuid.UUID]] = {}
self._cached_data: Dict[uuid.UUID, Any] = {}
self._load_configs: Dict[uuid.UUID, Optional[LoadConfig]] = {}
# initialize special values
special_value_cls = PythonClass.from_class(SpecialValue)
self._not_set_value: Value = Value(
value_id=NOT_SET_VALUE_ID,
kiara_id=self._kiara.id,
value_schema=ValueSchema(
type="none",
default=SpecialValue.NOT_SET,
is_constant=True,
doc="Special value, indicating a field is not set.", # type: ignore
),
value_status=ValueStatus.NOT_SET,
value_size=0,
value_hash=INVALID_HASH_MARKER,
pedigree=ORPHAN,
pedigree_output_name="__void__",
data_type_class=special_value_cls,
)
self._cached_data[NOT_SET_VALUE_ID] = SpecialValue.NOT_SET
self._none_value: Value = Value(
value_id=NONE_VALUE_ID,
kiara_id=self._kiara.id,
value_schema=ValueSchema(
type="special_type",
default=SpecialValue.NO_VALUE,
is_constant=True,
doc="Special value, indicating a field is set with a 'none' value.", # type: ignore
),
value_status=ValueStatus.NONE,
value_size=0,
value_hash=-2,
pedigree=ORPHAN,
pedigree_output_name="__void__",
data_type_class=special_value_cls,
)
self._cached_data[NONE_VALUE_ID] = SpecialValue.NO_VALUE
@property
def kiara_id(self) -> uuid.UUID:
return self._kiara.id
@property
def NOT_SET_VALUE(self) -> Value:
return self._not_set_value
@property
def NONE_VALUE(self) -> Value:
return self._none_value
def retrieve_all_available_value_ids(self) -> Set[uuid.UUID]:
result: Set[uuid.UUID] = set()
for store in self._data_archives.values():
ids = store.value_ids
result.update(ids)
return result
def register_data_archive(
self,
archive: DataArchive,
alias: str = None,
set_as_default_store: Optional[bool] = None,
):
data_store_id = archive.register_archive(kiara=self._kiara)
if alias is None:
alias = str(data_store_id)
if alias in self._data_archives.keys():
raise Exception(f"Can't add store, alias '{alias}' already registered.")
self._data_archives[alias] = archive
is_store = False
is_default_store = False
if isinstance(archive, DataStore):
is_store = True
if set_as_default_store and self._default_data_store is not None:
raise Exception(
f"Can't set data store '{alias}' as default store: default store already set."
)
if self._default_data_store is None or set_as_default_store:
is_default_store = True
self._default_data_store = alias
event = DataArchiveAddedEvent.construct(
kiara_id=self._kiara.id,
data_archive_id=archive.archive_id,
data_archive_alias=alias,
is_store=is_store,
is_default_store=is_default_store,
)
self._event_callback(event)
@property
def default_data_store(self) -> str:
if self._default_data_store is None:
raise Exception("No default data store set.")
return self._default_data_store
@property
def data_archives(self) -> Mapping[str, DataArchive]:
return self._data_archives
def get_archive(self, store_id: Optional[str] = None) -> DataArchive:
if store_id is None:
store_id = self.default_data_store
if store_id is None:
raise Exception("Can't retrieve default data archive, none set (yet).")
return self._data_archives[store_id]
def find_store_id_for_value(self, value_id: uuid.UUID) -> Optional[str]:
if value_id in self._value_archive_lookup_map.keys():
return self._value_archive_lookup_map[value_id]
matches = []
for store_id, store in self.data_archives.items():
match = store.has_value(value_id=value_id)
if match:
matches.append(store_id)
if len(matches) == 0:
return None
elif len(matches) > 1:
raise Exception(
f"Found value with id '{value_id}' in multiple archives, this is not supported (yet): {matches}"
)
self._value_archive_lookup_map[value_id] = matches[0]
return matches[0]
def get_value(self, value_id: Union[uuid.UUID, Value, str]) -> Value:
_value_id = None
if not isinstance(value_id, uuid.UUID):
# fallbacks for common mistakes, this should error out if not a Value or string.
if hasattr(value_id, "value_id"):
_value_id: Optional[uuid.UUID] = value_id.value_id # type: ignore
else:
try:
_value_id = uuid.UUID(
value_id # type: ignore
) # this should fail if not string or wrong string format
except ValueError:
if not isinstance(value_id, str):
raise Exception(
f"Can't retrieve value for '{value_id}': invalid type '{type(value_id)}'."
)
if ":" not in value_id:
raise Exception(
f"Can't retrieve value for '{value_id}': can't determine reference type."
)
ref_type, rest = value_id.split(":", maxsplit=1)
if ref_type == "value":
_value_id = uuid.UUID(rest)
elif ref_type == "alias":
_value_id = self._kiara.alias_registry.find_value_id_for_alias(
alias=rest
)
if _value_id is None:
raise Exception(
f"Can't retrive value for alias '{rest}': no such alias registered."
)
else:
raise Exception(
f"Can't retrieve value for '{value_id}': invalid reference type '{ref_type}'."
)
else:
_value_id = value_id
assert _value_id is not None
if _value_id in self._registered_values.keys():
value = self._registered_values[_value_id]
return value
matches = []
for store_id, store in self.data_archives.items():
match = store.has_value(value_id=_value_id)
if match:
matches.append(store_id)
if len(matches) == 0:
raise Exception(f"No value registered with id: {value_id}")
elif len(matches) > 1:
raise Exception(
f"Found value with id '{value_id}' in multiple archives, this is not supported (yet): {matches}"
)
self._value_archive_lookup_map[_value_id] = matches[0]
stored_value = self.get_archive(matches[0]).retrieve_value(value_id=_value_id)
stored_value._set_registry(self)
stored_value._is_stored = True
self._registered_values[_value_id] = stored_value
return self._registered_values[_value_id]
def store_value(
self,
value: Union[Value, uuid.UUID],
store_id: Optional[str] = None,
skip_if_exists: bool = True,
):
if store_id is None:
store_id = self.default_data_store
if isinstance(value, uuid.UUID):
value = self.get_value(value)
store: DataStore = self.get_archive(store_id=store_id) # type: ignore
if not isinstance(store, DataStore):
raise Exception(f"Can't store value into store '{store_id}': not writable.")
# make sure all property values are available
if value.pedigree != ORPHAN:
for value_id in value.pedigree.inputs.values():
self.store_value(value=value_id, store_id=store_id, skip_if_exists=True)
if not store.has_value(value.value_id) or not skip_if_exists:
event = ValuePreStoreEvent.construct(kiara_id=self._kiara.id, value=value)
self._event_callback(event)
load_config = store.store_value(value)
value._is_stored = True
self._value_archive_lookup_map[value.value_id] = store_id
self._load_configs[value.value_id] = load_config
property_values = value.property_values
for property, property_value in property_values.items():
self.store_value(
value=property_value, store_id=store_id, skip_if_exists=True
)
store_event = ValueStoredEvent.construct(kiara_id=self._kiara.id, value=value)
self._event_callback(store_event)
def find_values_for_hash(
self, value_hash: int, data_type_name: Optional[str] = None
) -> Set[Value]:
if data_type_name:
raise NotImplementedError()
stored = self._values_by_hash.get(value_hash, None)
if stored is None:
matches: Dict[uuid.UUID, List[str]] = {}
for store_id, store in self.data_archives.items():
value_ids = store.find_values_with_hash(
value_hash=value_hash, data_type_name=data_type_name
)
for v_id in value_ids:
matches.setdefault(v_id, []).append(store_id)
stored = set()
for v_id, store_ids in matches.items():
if len(store_ids) > 1:
raise Exception(
f"Found multiple stores for value id '{v_id}', this is not supported (yet)."
)
self._value_archive_lookup_map[v_id] = store_ids[0]
stored.add(v_id)
if stored:
self._values_by_hash[value_hash] = stored
return set((self.get_value(value_id=v_id) for v_id in stored))
def find_destinies_for_value(
self, value_id: uuid.UUID, alias_filter: str = None
) -> Mapping[str, uuid.UUID]:
if alias_filter:
raise NotImplementedError()
all_destinies: Dict[str, uuid.UUID] = {}
for archive_id, archive in self._data_archives.items():
destinies: Optional[
Mapping[str, uuid.UUID]
] = archive.find_destinies_for_value(
value_id=value_id, alias_filter=alias_filter
)
if not destinies:
continue
for k, v in destinies.items():
if k in all_destinies.keys():
raise Exception(f"Duplicate destiny '{k}' for value '{value_id}'.")
all_destinies[k] = v
return all_destinies
def register_data(
self,
data: Any,
schema: Optional[ValueSchema] = None,
pedigree: Optional[ValuePedigree] = None,
pedigree_output_name: str = None,
reuse_existing: bool = True,
) -> Value:
value, newly_created = self._create_value(
data=data,
schema=schema,
pedigree=pedigree,
pedigree_output_name=pedigree_output_name,
reuse_existing=reuse_existing,
)
if newly_created:
self._values_by_hash.setdefault(value.value_hash, set()).add(value.value_id)
self._registered_values[value.value_id] = value
self._cached_data[value.value_id] = data
return value
def _find_existing_value(
self, data: Any, schema: Optional[ValueSchema]
) -> Tuple[
Optional[Value], Optional[DataType], Optional[ValueStatus], Optional[int]
]:
if schema is None:
raise NotImplementedError()
if isinstance(data, Value):
if data.value_id in self._registered_values.keys():
return (data, None, None, None)
raise NotImplementedError("Importing values not supported (yet).")
# self._registered_values[data.value_id] = data
# return data
try:
value = self.get_value(value_id=data)
return (value, None, None, None)
except Exception:
# TODO: differentiate between 'value not found' and other type of errors
pass
# no obvious matches, so we try to find data that has the same hash
data_type = self._kiara.type_registry.retrieve_data_type(
data_type_name=schema.type, data_type_config=schema.type_config
)
if data == SpecialValue.NOT_SET:
status = ValueStatus.NOT_SET
value_hash = INVALID_HASH_MARKER
elif data == SpecialValue.NO_VALUE:
status = ValueStatus.NONE
value_hash = INVALID_HASH_MARKER
else:
data, status, value_hash = data_type._pre_examine_data(
data=data, schema=schema
)
existing_value: Optional[Value] = None
if value_hash != INVALID_HASH_MARKER:
existing = self.find_values_for_hash(value_hash=value_hash)
if existing:
if len(existing) == 1:
existing_value = next(iter(existing))
else:
skalars = []
for v in existing:
if v.data_type.characteristics.is_scalar:
skalars.append(v)
if len(skalars) == 1:
existing_value = skalars[0]
elif skalars:
orphans = []
for v in skalars:
if v.pedigree == ORPHAN:
orphans.append(v)
if len(orphans) == 1:
existing_value = orphans[0]
if existing_value is not None:
self._load_configs[existing_value.value_id] = None
return (existing_value, data_type, status, value_hash)
return (None, data_type, status, value_hash)
def _create_value(
self,
data: Any,
schema: Optional[ValueSchema] = None,
pedigree: Optional[ValuePedigree] = None,
pedigree_output_name: str = None,
reuse_existing: bool = True,
data_is_value_reference: Optional[bool] = None,
) -> Tuple[Value, bool]:
"""Create a new value, or return an existing one that matches the incoming data or reference.
Arguments:
data: the (raw) data, or a reference to an existing value
Returns:
a tuple containing of the value object, and a boolean indicating whether the value was newly created (True), or already existing (False)
"""
if schema is None:
raise NotImplementedError()
if schema.type not in self._kiara.data_type_names:
raise Exception(
f"Can't register data of type '{schema.type}': type not registered. Available types: {', '.join(self._kiara.data_type_names)}"
)
data_type: Optional[DataType] = None
status: Optional[ValueStatus] = None
value_hash: Optional[int] = None
if reuse_existing:
existing, data_type, status, value_hash = self._find_existing_value(
data=data, schema=schema
)
if existing is not None:
# TODO: check pedigree
return (existing, False)
if pedigree is None:
pedigree = ORPHAN
if pedigree_output_name is None:
if pedigree == ORPHAN:
pedigree_output_name = ORPHAN_PEDIGREE_OUTPUT_NAME
else:
raise NotImplementedError()
v_id = ID_REGISTRY.generate(
type="value", kiara_id=self._kiara.id, pre_registered=False
)
if data_type is None or status is None or value_hash is None:
data_type = self._kiara.type_registry.retrieve_data_type(
data_type_name=schema.type, data_type_config=schema.type_config
)
if data == SpecialValue.NOT_SET:
status = ValueStatus.NOT_SET
value_hash = INVALID_HASH_MARKER
elif data == SpecialValue.NO_VALUE:
status = ValueStatus.NONE
value_hash = INVALID_HASH_MARKER
else:
data, status, value_hash = data_type._pre_examine_data(
data=data, schema=schema
)
value, data = data_type.assemble_value(
value_id=v_id,
data=data,
schema=schema,
status=status,
value_hash=value_hash,
pedigree=pedigree,
kiara_id=self._kiara.id,
pedigree_output_name=pedigree_output_name,
)
ID_REGISTRY.update_metadata(v_id, obj=value)
value._data_registry = self
event = ValueCreatedEvent(kiara_id=self._kiara.id, value=value)
self._event_callback(event)
return (value, True)
def retrieve_load_config(self, value_id: uuid.UUID) -> Optional[LoadConfig]:
if (
value_id in self._load_configs.keys()
and self._load_configs[value_id] is not None
):
load_config = self._load_configs[value_id]
else:
# now, the value_store map should contain this value_id
store_id = self.find_store_id_for_value(value_id=value_id)
if store_id is None:
return None
store = self.get_archive(store_id)
assert value_id in self._registered_values.keys()
# self.get_value(value_id=value_id)
load_config = store.retrieve_load_config(value=value_id)
self._load_configs[value_id] = load_config
return load_config
def _retrieve_bytes(
self, chunk_id: str, as_bytes: bool = True
) -> Union[str, bytes]:
# TODO: support multiple stores
return self.get_archive().retrieve_bytes(chunk_id=chunk_id, as_bytes=as_bytes)
def retrieve_value_data(self, value_id: uuid.UUID) -> Any:
if value_id in self._cached_data.keys():
return self._cached_data[value_id]
load_config = self.retrieve_load_config(value_id=value_id)
if load_config is None:
raise Exception(
f"Load config for value '{value_id}' is 'None', this is most likely a bug."
)
data = self._load_data_from_load_config(
load_config=load_config, value_id=value_id
)
self._cached_data[value_id] = data
return data
def _provision_bytes(
self, load_config: LoadConfig
) -> Tuple[Optional[BytesStructure], Optional[Path]]:
provisioning_strategy = load_config.provisioning_strategy
if provisioning_strategy == ByteProvisioningStrategy.INLINE:
return None, None
assert load_config.bytes_map is not None
bytes_structure_map: Dict[str, List[Union[str, bytes]]] = {}
path: Optional[Path] = None
if provisioning_strategy == ByteProvisioningStrategy.BYTES:
for field, chunk_ids in load_config.bytes_map.chunk_id_map.items():
for chunk_id in chunk_ids:
bytes_structure_map.setdefault(field, []).append(
self._retrieve_bytes(chunk_id=chunk_id, as_bytes=True)
)
elif provisioning_strategy == ByteProvisioningStrategy.FILE_PATH_MAP:
for field, chunk_ids in load_config.bytes_map.chunk_id_map.items():
for chunk_id in chunk_ids:
bytes_structure_map.setdefault(field, []).append(
self._retrieve_bytes(chunk_id=chunk_id, as_bytes=False)
)
elif provisioning_strategy == ByteProvisioningStrategy.LINK_FOLDER:
temp_f = tempfile.mkdtemp()
def cleanup():
logger.debug("deleting.temp_provisioning_folder", path=temp_f)
shutil.rmtree(temp_f, ignore_errors=True)
atexit.register(cleanup)
root_dir = Path(temp_f)
for rel_path, chunk_ids in load_config.bytes_map.chunk_id_map.items():
if len(chunk_ids) != 1:
raise Exception(
f"Multiple chunks for file '{rel_path}' provided, this is currently not supported when using the 'LINK_FOLDER' provisioning strategy."
)
chunk_id = chunk_ids[0]
byte_ref: str = self._retrieve_bytes(chunk_id=chunk_id, as_bytes=False) # type: ignore
path = Path(byte_ref)
link = root_dir / rel_path
link.parent.mkdir(parents=True, exist_ok=True)
link.symlink_to(path)
bytes_structure_map[rel_path] = [link.as_posix()]
path = root_dir
elif provisioning_strategy == ByteProvisioningStrategy.COPIED_FOLDER:
raise NotImplementedError("xxxxxxxx")
bytes_structure = BytesStructure.construct(
data_type=load_config.bytes_map.data_type,
data_type_config=load_config.bytes_map.data_type_config,
chunk_map=bytes_structure_map,
)
return bytes_structure, path
def _load_data_from_load_config(
self, load_config: LoadConfig, value_id: uuid.UUID
) -> Any:
logger.debug("value.load", module=load_config.module_type)
# TODO: check whether modules and value types are available
inputs: Dict[str, Any] = dict(load_config.inputs)
bytes_structure, path = self._provision_bytes(load_config=load_config)
if load_config.provisioning_strategy == ByteProvisioningStrategy.INLINE:
if "inline_data" in load_config.inputs.keys():
assert load_config.inputs["inline_data"] == LOAD_CONFIG_PLACEHOLDER
inputs["inline_data"] = load_config.inline_data
elif load_config.provisioning_strategy == ByteProvisioningStrategy.BYTES:
if "inline_data" in load_config.inputs.keys():
assert load_config.inputs["inline_data"] == LOAD_CONFIG_PLACEHOLDER
inputs["inline_data"] = load_config.inline_data
if "bytes_structure" in load_config.inputs.keys():
assert load_config.inputs["bytes_structure"] == LOAD_CONFIG_PLACEHOLDER
inputs["bytes_structure"] = bytes_structure
elif (
load_config.provisioning_strategy == ByteProvisioningStrategy.FILE_PATH_MAP
):
if "inline_data" in load_config.inputs.keys():
assert load_config.inputs["inline_data"] == LOAD_CONFIG_PLACEHOLDER
inputs["inline_data"] = load_config.inline_data
if "bytes_structure" in load_config.inputs.keys():
assert load_config.inputs["bytes_structure"] == LOAD_CONFIG_PLACEHOLDER
inputs["bytes_structure"] = bytes_structure
elif load_config.provisioning_strategy == ByteProvisioningStrategy.LINK_FOLDER:
if "inline_data" in load_config.inputs.keys():
assert load_config.inputs["inline_data"] == LOAD_CONFIG_PLACEHOLDER
inputs["inline_data"] = load_config.inline_data
if "bytes_structure" in load_config.inputs.keys():
assert load_config.inputs["bytes_structure"] == LOAD_CONFIG_PLACEHOLDER
inputs["bytes_structure"] = bytes_structure
if "path" in load_config.inputs.keys():
assert path is not None
assert load_config.inputs["path"] == LOAD_CONFIG_PLACEHOLDER
inputs["path"] = path.as_posix()
elif (
load_config.provisioning_strategy == ByteProvisioningStrategy.COPIED_FOLDER
):
raise NotImplementedError()
try:
job_config = self._kiara.job_registry.prepare_job_config(
manifest=load_config, inputs=inputs
)
except JobConfigException:
if is_debug():
import traceback
traceback.print_exc()
value = self.get_value(value_id=value_id)
return UnloadableData(value=value, load_config=load_config)
job_id = self._kiara.job_registry.execute_job(job_config=job_config)
result = self._kiara.job_registry.retrieve_result(job_id=job_id)
# data = result.get_value_data(load_config.output_name)
result_value = result.get_value_obj(field_name=load_config.output_name)
return result_value.data
def load_values(self, values: Mapping[str, Optional[uuid.UUID]]) -> ValueMap:
value_items = {}
schemas = {}
for field_name, value_id in values.items():
if value_id is None:
value_id = NONE_VALUE_ID
value = self.get_value(value_id=value_id)
value_items[field_name] = value
schemas[field_name] = value.value_schema
return ValueMapReadOnly(value_items=value_items, values_schema=schemas)
def load_data(self, values: Mapping[str, Optional[uuid.UUID]]) -> Mapping[str, Any]:
result_values = self.load_values(values=values)
return {k: v.data for k, v in result_values.items()}
def create_valueset(
self, data: Mapping[str, Any], schema: Mapping[str, ValueSchema]
) -> ValueMap:
"""Extract a set of [Value][kiara.data.values.Value] from Python data and ValueSchemas."""
input_details = {}
for input_name, value_schema in schema.items():
input_details[input_name] = {"schema": value_schema}
leftover = set(data.keys())
leftover.difference_update(input_details.keys())
if leftover:
if not STRICT_CHECKS:
log_message("unused.inputs", input_names=leftover)
else:
raise Exception(
f"Can't create values instance, inputs contain unused/invalid fields: {', '.join(leftover)}"
)
values = {}
failed = {}
for input_name, details in input_details.items():
value_schema = details["schema"]
if input_name not in data.keys():
value_data = SpecialValue.NOT_SET
elif data[input_name] is None:
value_data = SpecialValue.NO_VALUE
else:
value_data = data[input_name]
try:
value = self.register_data(
data=value_data, schema=value_schema, reuse_existing=True
)
# value = self.retrieve_or_create_value(
# value_data, value_schema=value_schema
# )
values[input_name] = value
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
failed[input_name] = e
if failed:
msg = []
for k, v in failed.items():
_v = str(v)
if not str(v):
_v = type(v).__name__
msg.append(f"{k}: {_v}")
raise InvalidValuesException(
msg=f"Can't create values instance: {', '.join(msg)}",
invalid_values={k: str(v) for k, v in failed.items()},
)
return ValueMapReadOnly(value_items=values, values_schema=schema) # type: ignore
def create_renderable(self, **config: Any) -> RenderableType:
"""Create a renderable for this module configuration."""
from kiara.utils.output import create_renderable_from_values
all_values = {str(i): v for i, v in self._registered_values.items()}
table = create_renderable_from_values(values=all_values, config=config)
return table
def render_data(
self,
value_id: uuid.UUID,
target_type="terminal_renderable",
**render_config: Any,
) -> Any:
assert isinstance(value_id, uuid.UUID)
return render_data(
kiara=self._kiara,
value_id=value_id,
target_type=target_type,
**render_config,
)
NONE_VALUE: Value
property
readonly
¶NOT_SET_VALUE: Value
property
readonly
¶data_archives: Mapping[str, kiara.registries.data.data_store.DataArchive]
property
readonly
¶default_data_store: str
property
readonly
¶kiara_id: UUID
property
readonly
¶create_renderable(self, **config)
¶Create a renderable for this module configuration.
Source code in kiara/registries/data/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
"""Create a renderable for this module configuration."""
from kiara.utils.output import create_renderable_from_values
all_values = {str(i): v for i, v in self._registered_values.items()}
table = create_renderable_from_values(values=all_values, config=config)
return table
create_valueset(self, data, schema)
¶Extract a set of [Value][kiara.data.values.Value] from Python data and ValueSchemas.
Source code in kiara/registries/data/__init__.py
def create_valueset(
self, data: Mapping[str, Any], schema: Mapping[str, ValueSchema]
) -> ValueMap:
"""Extract a set of [Value][kiara.data.values.Value] from Python data and ValueSchemas."""
input_details = {}
for input_name, value_schema in schema.items():
input_details[input_name] = {"schema": value_schema}
leftover = set(data.keys())
leftover.difference_update(input_details.keys())
if leftover:
if not STRICT_CHECKS:
log_message("unused.inputs", input_names=leftover)
else:
raise Exception(
f"Can't create values instance, inputs contain unused/invalid fields: {', '.join(leftover)}"
)
values = {}
failed = {}
for input_name, details in input_details.items():
value_schema = details["schema"]
if input_name not in data.keys():
value_data = SpecialValue.NOT_SET
elif data[input_name] is None:
value_data = SpecialValue.NO_VALUE
else:
value_data = data[input_name]
try:
value = self.register_data(
data=value_data, schema=value_schema, reuse_existing=True
)
# value = self.retrieve_or_create_value(
# value_data, value_schema=value_schema
# )
values[input_name] = value
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
failed[input_name] = e
if failed:
msg = []
for k, v in failed.items():
_v = str(v)
if not str(v):
_v = type(v).__name__
msg.append(f"{k}: {_v}")
raise InvalidValuesException(
msg=f"Can't create values instance: {', '.join(msg)}",
invalid_values={k: str(v) for k, v in failed.items()},
)
return ValueMapReadOnly(value_items=values, values_schema=schema) # type: ignore
find_destinies_for_value(self, value_id, alias_filter=None)
¶Source code in kiara/registries/data/__init__.py
def find_destinies_for_value(
self, value_id: uuid.UUID, alias_filter: str = None
) -> Mapping[str, uuid.UUID]:
if alias_filter:
raise NotImplementedError()
all_destinies: Dict[str, uuid.UUID] = {}
for archive_id, archive in self._data_archives.items():
destinies: Optional[
Mapping[str, uuid.UUID]
] = archive.find_destinies_for_value(
value_id=value_id, alias_filter=alias_filter
)
if not destinies:
continue
for k, v in destinies.items():
if k in all_destinies.keys():
raise Exception(f"Duplicate destiny '{k}' for value '{value_id}'.")
all_destinies[k] = v
return all_destinies
find_store_id_for_value(self, value_id)
¶Source code in kiara/registries/data/__init__.py
def find_store_id_for_value(self, value_id: uuid.UUID) -> Optional[str]:
if value_id in self._value_archive_lookup_map.keys():
return self._value_archive_lookup_map[value_id]
matches = []
for store_id, store in self.data_archives.items():
match = store.has_value(value_id=value_id)
if match:
matches.append(store_id)
if len(matches) == 0:
return None
elif len(matches) > 1:
raise Exception(
f"Found value with id '{value_id}' in multiple archives, this is not supported (yet): {matches}"
)
self._value_archive_lookup_map[value_id] = matches[0]
return matches[0]
find_values_for_hash(self, value_hash, data_type_name=None)
¶Source code in kiara/registries/data/__init__.py
def find_values_for_hash(
self, value_hash: int, data_type_name: Optional[str] = None
) -> Set[Value]:
if data_type_name:
raise NotImplementedError()
stored = self._values_by_hash.get(value_hash, None)
if stored is None:
matches: Dict[uuid.UUID, List[str]] = {}
for store_id, store in self.data_archives.items():
value_ids = store.find_values_with_hash(
value_hash=value_hash, data_type_name=data_type_name
)
for v_id in value_ids:
matches.setdefault(v_id, []).append(store_id)
stored = set()
for v_id, store_ids in matches.items():
if len(store_ids) > 1:
raise Exception(
f"Found multiple stores for value id '{v_id}', this is not supported (yet)."
)
self._value_archive_lookup_map[v_id] = store_ids[0]
stored.add(v_id)
if stored:
self._values_by_hash[value_hash] = stored
return set((self.get_value(value_id=v_id) for v_id in stored))
get_archive(self, store_id=None)
¶Source code in kiara/registries/data/__init__.py
def get_archive(self, store_id: Optional[str] = None) -> DataArchive:
if store_id is None:
store_id = self.default_data_store
if store_id is None:
raise Exception("Can't retrieve default data archive, none set (yet).")
return self._data_archives[store_id]
get_value(self, value_id)
¶Source code in kiara/registries/data/__init__.py
def get_value(self, value_id: Union[uuid.UUID, Value, str]) -> Value:
_value_id = None
if not isinstance(value_id, uuid.UUID):
# fallbacks for common mistakes, this should error out if not a Value or string.
if hasattr(value_id, "value_id"):
_value_id: Optional[uuid.UUID] = value_id.value_id # type: ignore
else:
try:
_value_id = uuid.UUID(
value_id # type: ignore
) # this should fail if not string or wrong string format
except ValueError:
if not isinstance(value_id, str):
raise Exception(
f"Can't retrieve value for '{value_id}': invalid type '{type(value_id)}'."
)
if ":" not in value_id:
raise Exception(
f"Can't retrieve value for '{value_id}': can't determine reference type."
)
ref_type, rest = value_id.split(":", maxsplit=1)
if ref_type == "value":
_value_id = uuid.UUID(rest)
elif ref_type == "alias":
_value_id = self._kiara.alias_registry.find_value_id_for_alias(
alias=rest
)
if _value_id is None:
raise Exception(
f"Can't retrive value for alias '{rest}': no such alias registered."
)
else:
raise Exception(
f"Can't retrieve value for '{value_id}': invalid reference type '{ref_type}'."
)
else:
_value_id = value_id
assert _value_id is not None
if _value_id in self._registered_values.keys():
value = self._registered_values[_value_id]
return value
matches = []
for store_id, store in self.data_archives.items():
match = store.has_value(value_id=_value_id)
if match:
matches.append(store_id)
if len(matches) == 0:
raise Exception(f"No value registered with id: {value_id}")
elif len(matches) > 1:
raise Exception(
f"Found value with id '{value_id}' in multiple archives, this is not supported (yet): {matches}"
)
self._value_archive_lookup_map[_value_id] = matches[0]
stored_value = self.get_archive(matches[0]).retrieve_value(value_id=_value_id)
stored_value._set_registry(self)
stored_value._is_stored = True
self._registered_values[_value_id] = stored_value
return self._registered_values[_value_id]
load_data(self, values)
¶Source code in kiara/registries/data/__init__.py
def load_data(self, values: Mapping[str, Optional[uuid.UUID]]) -> Mapping[str, Any]:
result_values = self.load_values(values=values)
return {k: v.data for k, v in result_values.items()}
load_values(self, values)
¶Source code in kiara/registries/data/__init__.py
def load_values(self, values: Mapping[str, Optional[uuid.UUID]]) -> ValueMap:
value_items = {}
schemas = {}
for field_name, value_id in values.items():
if value_id is None:
value_id = NONE_VALUE_ID
value = self.get_value(value_id=value_id)
value_items[field_name] = value
schemas[field_name] = value.value_schema
return ValueMapReadOnly(value_items=value_items, values_schema=schemas)
register_data(self, data, schema=None, pedigree=None, pedigree_output_name=None, reuse_existing=True)
¶Source code in kiara/registries/data/__init__.py
def register_data(
self,
data: Any,
schema: Optional[ValueSchema] = None,
pedigree: Optional[ValuePedigree] = None,
pedigree_output_name: str = None,
reuse_existing: bool = True,
) -> Value:
value, newly_created = self._create_value(
data=data,
schema=schema,
pedigree=pedigree,
pedigree_output_name=pedigree_output_name,
reuse_existing=reuse_existing,
)
if newly_created:
self._values_by_hash.setdefault(value.value_hash, set()).add(value.value_id)
self._registered_values[value.value_id] = value
self._cached_data[value.value_id] = data
return value
register_data_archive(self, archive, alias=None, set_as_default_store=None)
¶Source code in kiara/registries/data/__init__.py
def register_data_archive(
self,
archive: DataArchive,
alias: str = None,
set_as_default_store: Optional[bool] = None,
):
data_store_id = archive.register_archive(kiara=self._kiara)
if alias is None:
alias = str(data_store_id)
if alias in self._data_archives.keys():
raise Exception(f"Can't add store, alias '{alias}' already registered.")
self._data_archives[alias] = archive
is_store = False
is_default_store = False
if isinstance(archive, DataStore):
is_store = True
if set_as_default_store and self._default_data_store is not None:
raise Exception(
f"Can't set data store '{alias}' as default store: default store already set."
)
if self._default_data_store is None or set_as_default_store:
is_default_store = True
self._default_data_store = alias
event = DataArchiveAddedEvent.construct(
kiara_id=self._kiara.id,
data_archive_id=archive.archive_id,
data_archive_alias=alias,
is_store=is_store,
is_default_store=is_default_store,
)
self._event_callback(event)
render_data(self, value_id, target_type='terminal_renderable', **render_config)
¶Source code in kiara/registries/data/__init__.py
def render_data(
self,
value_id: uuid.UUID,
target_type="terminal_renderable",
**render_config: Any,
) -> Any:
assert isinstance(value_id, uuid.UUID)
return render_data(
kiara=self._kiara,
value_id=value_id,
target_type=target_type,
**render_config,
)
retrieve_all_available_value_ids(self)
¶Source code in kiara/registries/data/__init__.py
def retrieve_all_available_value_ids(self) -> Set[uuid.UUID]:
result: Set[uuid.UUID] = set()
for store in self._data_archives.values():
ids = store.value_ids
result.update(ids)
return result
retrieve_load_config(self, value_id)
¶Source code in kiara/registries/data/__init__.py
def retrieve_load_config(self, value_id: uuid.UUID) -> Optional[LoadConfig]:
if (
value_id in self._load_configs.keys()
and self._load_configs[value_id] is not None
):
load_config = self._load_configs[value_id]
else:
# now, the value_store map should contain this value_id
store_id = self.find_store_id_for_value(value_id=value_id)
if store_id is None:
return None
store = self.get_archive(store_id)
assert value_id in self._registered_values.keys()
# self.get_value(value_id=value_id)
load_config = store.retrieve_load_config(value=value_id)
self._load_configs[value_id] = load_config
return load_config
retrieve_value_data(self, value_id)
¶Source code in kiara/registries/data/__init__.py
def retrieve_value_data(self, value_id: uuid.UUID) -> Any:
if value_id in self._cached_data.keys():
return self._cached_data[value_id]
load_config = self.retrieve_load_config(value_id=value_id)
if load_config is None:
raise Exception(
f"Load config for value '{value_id}' is 'None', this is most likely a bug."
)
data = self._load_data_from_load_config(
load_config=load_config, value_id=value_id
)
self._cached_data[value_id] = data
return data
store_value(self, value, store_id=None, skip_if_exists=True)
¶Source code in kiara/registries/data/__init__.py
def store_value(
self,
value: Union[Value, uuid.UUID],
store_id: Optional[str] = None,
skip_if_exists: bool = True,
):
if store_id is None:
store_id = self.default_data_store
if isinstance(value, uuid.UUID):
value = self.get_value(value)
store: DataStore = self.get_archive(store_id=store_id) # type: ignore
if not isinstance(store, DataStore):
raise Exception(f"Can't store value into store '{store_id}': not writable.")
# make sure all property values are available
if value.pedigree != ORPHAN:
for value_id in value.pedigree.inputs.values():
self.store_value(value=value_id, store_id=store_id, skip_if_exists=True)
if not store.has_value(value.value_id) or not skip_if_exists:
event = ValuePreStoreEvent.construct(kiara_id=self._kiara.id, value=value)
self._event_callback(event)
load_config = store.store_value(value)
value._is_stored = True
self._value_archive_lookup_map[value.value_id] = store_id
self._load_configs[value.value_id] = load_config
property_values = value.property_values
for property, property_value in property_values.items():
self.store_value(
value=property_value, store_id=store_id, skip_if_exists=True
)
store_event = ValueStoredEvent.construct(kiara_id=self._kiara.id, value=value)
self._event_callback(store_event)
Modules¶
data_store
special
¶logger
¶
BaseDataStore (DataStore)
¶Source code in kiara/registries/data/data_store/__init__.py
class BaseDataStore(DataStore):
@abc.abstractmethod
def _persist_bytes(self, bytes_structure: BytesStructure) -> BytesAliasStructure:
pass
@abc.abstractmethod
def _persist_load_config(self, value: Value, load_config: LoadConfig):
pass
@abc.abstractmethod
def _persist_value_details(self, value: Value):
pass
@abc.abstractmethod
def _persist_value_data(
self, value: Value
) -> Tuple[LoadConfig, Optional[BytesStructure]]:
pass
@abc.abstractmethod
def _persist_value_pedigree(self, value: Value):
"""Create an internal link from a value to its pedigree (and pedigree details).
This is so that the 'retrieve_job_record' can be used to prevent running the same job again, and the link of value
to the job that produced it is preserved.
"""
@abc.abstractmethod
def _persist_environment_details(
self, env_type: str, env_hash: int, env_data: Mapping[str, Any]
):
pass
@abc.abstractmethod
def _persist_destiny_backlinks(self, value: Value):
pass
def store_value(self, value: Value) -> LoadConfig:
logger.debug(
"store.value",
data_type=value.value_schema.type,
value_id=value.value_id,
value_hash=value.value_hash,
)
# first, persist environment information
for env_type, env_hash in value.pedigree.environments.items():
cached = self._env_cache.get(env_type, {}).get(env_hash, None)
if cached is not None:
continue
env = self.kiara_context.environment_registry.get_environment_for_hash(
env_hash
)
self.persist_environment(env)
# save the value data and metadata
load_config = self._persist_value(value)
self._load_config_cache[value.value_id] = load_config
self._value_cache[value.value_id] = value
self._value_hash_index.setdefault(value.value_hash, set()).add(value.value_id)
# now link the output values to the manifest
# then, make sure the manifest is persisted
self._persist_value_pedigree(value=value)
return load_config
def _persist_value(self, value: Value) -> LoadConfig:
# TODO: check if value id is already persisted?
load_config, bytes_structure = self._persist_value_data(value=value)
if not load_config:
raise Exception(
"Can't write load config, no load config returned when persisting value."
)
if not isinstance(load_config, LoadConfig):
raise Exception(
f"Can't write load config, invalid result type '{type(load_config)}' when persisting value."
)
if (
load_config.provisioning_strategy != ByteProvisioningStrategy.INLINE
and not bytes_structure
):
raise Exception(
"Can't write load config, no bytes structure returned when persisting value."
)
if bytes_structure and not isinstance(bytes_structure, BytesStructure):
raise Exception(
f"Can't write load config, invalid result bytes structure type '{type(bytes_structure)}' when persisting value."
)
if bytes_structure:
alias_structure = self._persist_bytes(bytes_structure=bytes_structure)
load_config.bytes_map = alias_structure
self._persist_load_config(value=value, load_config=load_config)
self._persist_value_details(value=value)
if value.destiny_backlinks:
self._persist_destiny_backlinks(value=value)
return load_config
def persist_environment(self, environment: RuntimeEnvironment):
"""Persist the specified environment.
The environment is stored as a dictionary, including it's schema, not as the actual Python model.
This is to make sure it can still be loaded later on, in case the Python model has changed in later versions.
"""
env_type = environment.get_environment_type_name()
env_hash = environment.model_data_hash
env = self._env_cache.get(env_type, {}).get(env_hash, None)
if env is not None:
return
env_data = environment.as_dict_with_schema()
self._persist_environment_details(
env_type=env_type, env_hash=env_hash, env_data=env_data
)
self._env_cache.setdefault(env_type, {})[env_hash] = env_data
def create_renderable(self, **config: Any) -> RenderableType:
"""Create a renderable for this module configuration."""
from kiara.utils.output import create_renderable_from_values
all_values = {}
for value_id in self.value_ids:
value = self.kiara_context.data_registry.get_value(value_id)
all_values[str(value_id)] = value
table = create_renderable_from_values(values=all_values, config=config)
return table
create_renderable(self, **config)
¶Create a renderable for this module configuration.
Source code in kiara/registries/data/data_store/__init__.py
def create_renderable(self, **config: Any) -> RenderableType:
"""Create a renderable for this module configuration."""
from kiara.utils.output import create_renderable_from_values
all_values = {}
for value_id in self.value_ids:
value = self.kiara_context.data_registry.get_value(value_id)
all_values[str(value_id)] = value
table = create_renderable_from_values(values=all_values, config=config)
return table
persist_environment(self, environment)
¶Persist the specified environment.
The environment is stored as a dictionary, including it's schema, not as the actual Python model. This is to make sure it can still be loaded later on, in case the Python model has changed in later versions.
Source code in kiara/registries/data/data_store/__init__.py
def persist_environment(self, environment: RuntimeEnvironment):
"""Persist the specified environment.
The environment is stored as a dictionary, including it's schema, not as the actual Python model.
This is to make sure it can still be loaded later on, in case the Python model has changed in later versions.
"""
env_type = environment.get_environment_type_name()
env_hash = environment.model_data_hash
env = self._env_cache.get(env_type, {}).get(env_hash, None)
if env is not None:
return
env_data = environment.as_dict_with_schema()
self._persist_environment_details(
env_type=env_type, env_hash=env_hash, env_data=env_data
)
self._env_cache.setdefault(env_type, {})[env_hash] = env_data
store_value(self, value)
¶"Store the value, its data and metadata into the store.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
value |
Value |
the value to persist |
required |
Returns:
| Type | Description |
|---|---|
LoadConfig |
the load config that is needed to retrieve the value data later |
Source code in kiara/registries/data/data_store/__init__.py
def store_value(self, value: Value) -> LoadConfig:
logger.debug(
"store.value",
data_type=value.value_schema.type,
value_id=value.value_id,
value_hash=value.value_hash,
)
# first, persist environment information
for env_type, env_hash in value.pedigree.environments.items():
cached = self._env_cache.get(env_type, {}).get(env_hash, None)
if cached is not None:
continue
env = self.kiara_context.environment_registry.get_environment_for_hash(
env_hash
)
self.persist_environment(env)
# save the value data and metadata
load_config = self._persist_value(value)
self._load_config_cache[value.value_id] = load_config
self._value_cache[value.value_id] = value
self._value_hash_index.setdefault(value.value_hash, set()).add(value.value_id)
# now link the output values to the manifest
# then, make sure the manifest is persisted
self._persist_value_pedigree(value=value)
return load_config
DataArchive (BaseArchive)
¶Source code in kiara/registries/data/data_store/__init__.py
class DataArchive(BaseArchive):
@classmethod
def supported_item_types(cls) -> Iterable[str]:
return ["data"]
def __init__(self, archive_id: uuid.UUID, config: ARCHIVE_CONFIG_CLS):
super().__init__(archive_id=archive_id, config=config)
self._env_cache: Dict[str, Dict[int, Mapping[str, Any]]] = {}
self._value_cache: Dict[uuid.UUID, Value] = {}
self._load_config_cache: Dict[uuid.UUID, LoadConfig] = {}
self._value_hash_index: Dict[int, Set[uuid.UUID]] = {}
def retrieve_load_config(self, value: Union[uuid.UUID, Value]) -> LoadConfig:
if isinstance(value, Value):
value_id: uuid.UUID = value.value_id
_value: Optional[Value] = value
else:
value_id = value
_value = None
if value_id in self._load_config_cache.keys():
return self._load_config_cache[value_id]
if _value is None:
_value = self.retrieve_value(value_id)
assert _value is not None
load_config = self._retrieve_load_config(value=_value)
self._load_config_cache[_value.value_id] = load_config
return load_config
@abc.abstractmethod
def _retrieve_load_config(self, value: Value) -> LoadConfig:
pass
def retrieve_value(self, value_id: uuid.UUID) -> Value:
cached = self._value_cache.get(value_id, None)
if cached is not None:
return cached
value_data = self._retrieve_value_details(value_id=value_id)
value_schema = ValueSchema(**value_data["value_schema"])
# data_type = self._kiara.get_value_type(
# data_type=value_schema.type, data_type_config=value_schema.type_config
# )
pedigree = ValuePedigree(**value_data["pedigree"])
value = Value(
value_id=value_data["value_id"],
kiara_id=self.kiara_context.id,
value_schema=value_schema,
value_status=value_data["value_status"],
value_size=value_data["value_size"],
value_hash=value_data["value_hash"],
pedigree=pedigree,
pedigree_output_name=value_data["pedigree_output_name"],
data_type_class=value_data["data_type_class"],
property_links=value_data["property_links"],
destiny_backlinks=value_data["destiny_backlinks"],
)
self._value_cache[value_id] = value
return self._value_cache[value_id]
@abc.abstractmethod
def _retrieve_value_details(self, value_id: uuid.UUID) -> Mapping[str, Any]:
pass
@property
def value_ids(self) -> Iterable[uuid.UUID]:
return self._retrieve_all_value_ids()
@abc.abstractmethod
def _retrieve_all_value_ids(
self, data_type_name: Optional[str] = None
) -> Iterable[uuid.UUID]:
pass
def has_value(self, value_id: uuid.UUID) -> bool:
"""Check whether the specific value_id is persisted in this data store.
Implementing classes are encouraged to override this method, and choose a suitable, implementation specific
way to quickly determine whether a value id is valid for this data store.
Arguments:
value_id: the id of the value to check.
Returns:
whether this data store contains the value with the specified id
"""
return value_id in self._retrieve_all_value_ids()
def retrieve_environment_details(
self, env_type: str, env_hash: int
) -> Mapping[str, Any]:
"""Retrieve the environment details with the specified type and hash.
The environment is stored by the data store as a dictionary, including it's schema, not as the actual Python model.
This is to make sure it can still be loaded later on, in case the Python model has changed in later versions.
"""
cached = self._env_cache.get(env_type, {}).get(env_hash, None)
if cached is not None:
return cached
env = self._retrieve_environment_details(env_type=env_type, env_hash=env_hash)
self._env_cache.setdefault(env_type, {})[env_hash] = env
return env
@abc.abstractmethod
def _retrieve_environment_details(
self, env_type: str, env_hash: int
) -> Mapping[str, Any]:
pass
def find_values_with_hash(
self,
value_hash: int,
value_size: Optional[int] = None,
data_type_name: Optional[str] = None,
) -> Set[uuid.UUID]:
if data_type_name is not None:
raise NotImplementedError()
if value_size is not None:
raise NotImplementedError()
if value_hash in self._value_hash_index.keys():
value_ids: Optional[Set[uuid.UUID]] = self._value_hash_index[value_hash]
else:
value_ids = self._find_values_with_hash(
value_hash=value_hash, data_type_name=data_type_name
)
if value_ids is None:
value_ids = set()
self._value_hash_index[value_hash] = value_ids
assert value_ids is not None
return value_ids
@abc.abstractmethod
def _find_values_with_hash(
self,
value_hash: int,
value_size: Optional[int] = None,
data_type_name: Optional[str] = None,
) -> Optional[Set[uuid.UUID]]:
pass
def find_destinies_for_value(
self, value_id: uuid.UUID, alias_filter: Optional[str] = None
) -> Optional[Mapping[str, uuid.UUID]]:
return self._find_destinies_for_value(
value_id=value_id, alias_filter=alias_filter
)
@abc.abstractmethod
def _find_destinies_for_value(
self, value_id: uuid.UUID, alias_filter: Optional[str] = None
) -> Optional[Mapping[str, uuid.UUID]]:
pass
@abc.abstractmethod
def retrieve_bytes(self, chunk_id: str, as_bytes: bool = True) -> Union[bytes, str]:
pass
# def retrieve_job_record(self, inputs_manifest: InputsManifest) -> Optional[JobRecord]:
# return self._retrieve_job_record(
# manifest_hash=inputs_manifest.manifest_hash, jobs_hash=inputs_manifest.jobs_hash
# )
#
# @abc.abstractmethod
# def _retrieve_job_record(
# self, manifest_hash: int, jobs_hash: int
# ) -> Optional[JobRecord]:
# pass
value_ids: Iterable[uuid.UUID]
property
readonly
¶find_destinies_for_value(self, value_id, alias_filter=None)
¶Source code in kiara/registries/data/data_store/__init__.py
def find_destinies_for_value(
self, value_id: uuid.UUID, alias_filter: Optional[str] = None
) -> Optional[Mapping[str, uuid.UUID]]:
return self._find_destinies_for_value(
value_id=value_id, alias_filter=alias_filter
)
find_values_with_hash(self, value_hash, value_size=None, data_type_name=None)
¶Source code in kiara/registries/data/data_store/__init__.py
def find_values_with_hash(
self,
value_hash: int,
value_size: Optional[int] = None,
data_type_name: Optional[str] = None,
) -> Set[uuid.UUID]:
if data_type_name is not None:
raise NotImplementedError()
if value_size is not None:
raise NotImplementedError()
if value_hash in self._value_hash_index.keys():
value_ids: Optional[Set[uuid.UUID]] = self._value_hash_index[value_hash]
else:
value_ids = self._find_values_with_hash(
value_hash=value_hash, data_type_name=data_type_name
)
if value_ids is None:
value_ids = set()
self._value_hash_index[value_hash] = value_ids
assert value_ids is not None
return value_ids
has_value(self, value_id)
¶Check whether the specific value_id is persisted in this data store.
Implementing classes are encouraged to override this method, and choose a suitable, implementation specific way to quickly determine whether a value id is valid for this data store.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
value_id |
UUID |
the id of the value to check. |
required |
Returns:
| Type | Description |
|---|---|
bool |
whether this data store contains the value with the specified id |
Source code in kiara/registries/data/data_store/__init__.py
def has_value(self, value_id: uuid.UUID) -> bool:
"""Check whether the specific value_id is persisted in this data store.
Implementing classes are encouraged to override this method, and choose a suitable, implementation specific
way to quickly determine whether a value id is valid for this data store.
Arguments:
value_id: the id of the value to check.
Returns:
whether this data store contains the value with the specified id
"""
return value_id in self._retrieve_all_value_ids()
retrieve_bytes(self, chunk_id, as_bytes=True)
¶Source code in kiara/registries/data/data_store/__init__.py
@abc.abstractmethod
def retrieve_bytes(self, chunk_id: str, as_bytes: bool = True) -> Union[bytes, str]:
pass
retrieve_environment_details(self, env_type, env_hash)
¶Retrieve the environment details with the specified type and hash.
The environment is stored by the data store as a dictionary, including it's schema, not as the actual Python model. This is to make sure it can still be loaded later on, in case the Python model has changed in later versions.
Source code in kiara/registries/data/data_store/__init__.py
def retrieve_environment_details(
self, env_type: str, env_hash: int
) -> Mapping[str, Any]:
"""Retrieve the environment details with the specified type and hash.
The environment is stored by the data store as a dictionary, including it's schema, not as the actual Python model.
This is to make sure it can still be loaded later on, in case the Python model has changed in later versions.
"""
cached = self._env_cache.get(env_type, {}).get(env_hash, None)
if cached is not None:
return cached
env = self._retrieve_environment_details(env_type=env_type, env_hash=env_hash)
self._env_cache.setdefault(env_type, {})[env_hash] = env
return env
retrieve_load_config(self, value)
¶Source code in kiara/registries/data/data_store/__init__.py
def retrieve_load_config(self, value: Union[uuid.UUID, Value]) -> LoadConfig:
if isinstance(value, Value):
value_id: uuid.UUID = value.value_id
_value: Optional[Value] = value
else:
value_id = value
_value = None
if value_id in self._load_config_cache.keys():
return self._load_config_cache[value_id]
if _value is None:
_value = self.retrieve_value(value_id)
assert _value is not None
load_config = self._retrieve_load_config(value=_value)
self._load_config_cache[_value.value_id] = load_config
return load_config
retrieve_value(self, value_id)
¶Source code in kiara/registries/data/data_store/__init__.py
def retrieve_value(self, value_id: uuid.UUID) -> Value:
cached = self._value_cache.get(value_id, None)
if cached is not None:
return cached
value_data = self._retrieve_value_details(value_id=value_id)
value_schema = ValueSchema(**value_data["value_schema"])
# data_type = self._kiara.get_value_type(
# data_type=value_schema.type, data_type_config=value_schema.type_config
# )
pedigree = ValuePedigree(**value_data["pedigree"])
value = Value(
value_id=value_data["value_id"],
kiara_id=self.kiara_context.id,
value_schema=value_schema,
value_status=value_data["value_status"],
value_size=value_data["value_size"],
value_hash=value_data["value_hash"],
pedigree=pedigree,
pedigree_output_name=value_data["pedigree_output_name"],
data_type_class=value_data["data_type_class"],
property_links=value_data["property_links"],
destiny_backlinks=value_data["destiny_backlinks"],
)
self._value_cache[value_id] = value
return self._value_cache[value_id]
supported_item_types()
classmethod
¶Source code in kiara/registries/data/data_store/__init__.py
@classmethod
def supported_item_types(cls) -> Iterable[str]:
return ["data"]
DataStore (DataArchive)
¶Source code in kiara/registries/data/data_store/__init__.py
class DataStore(DataArchive):
@classmethod
def is_writeable(cls) -> bool:
return True
@abc.abstractmethod
def store_value(self, value: Value) -> LoadConfig:
""" "Store the value, its data and metadata into the store.
Arguments:
value: the value to persist
Returns:
the load config that is needed to retrieve the value data later
"""
is_writeable()
classmethod
¶Source code in kiara/registries/data/data_store/__init__.py
@classmethod
def is_writeable(cls) -> bool:
return True
store_value(self, value)
¶"Store the value, its data and metadata into the store.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
value |
Value |
the value to persist |
required |
Returns:
| Type | Description |
|---|---|
LoadConfig |
the load config that is needed to retrieve the value data later |
Source code in kiara/registries/data/data_store/__init__.py
@abc.abstractmethod
def store_value(self, value: Value) -> LoadConfig:
""" "Store the value, its data and metadata into the store.
Arguments:
value: the value to persist
Returns:
the load config that is needed to retrieve the value data later
"""
filesystem_store
¶DEFAULT_HASHFS_DEPTH
¶DEFAULT_HASHFS_WIDTH
¶DEFAULT_HASH_FS_ALGORITHM
¶VALUE_DETAILS_FILE_NAME
¶logger
¶
EntityType (Enum)
¶
FileSystemDataArchive (DataArchive, JobArchive)
¶Data store that loads data from the local filesystem.
Source code in kiara/registries/data/data_store/filesystem_store.py
class FileSystemDataArchive(DataArchive, JobArchive):
"""Data store that loads data from the local filesystem."""
_archive_type_name = "filesystem_data_archive"
_config_cls = FileSystemArchiveConfig
# @classmethod
# def supported_item_types(cls) -> Iterable[str]:
#
# return ["data", "job_record"]
@classmethod
def is_writeable(cls) -> bool:
return False
def __init__(self, archive_id: uuid.UUID, config: FileSystemArchiveConfig):
DataArchive.__init__(self, archive_id=archive_id, config=config)
self._base_path: Optional[Path] = None
self._hashfs_path: Optional[Path] = None
self._hashfs: Optional[HashFS] = None
# def get_job_archive_id(self) -> uuid.UUID:
# return self._kiara.id
@property
def data_store_path(self) -> Path:
if self._base_path is not None:
return self._base_path
self._base_path = Path(self.config.base_path) / str(self.archive_id)
self._base_path.mkdir(parents=True, exist_ok=True)
return self._base_path
@property
def hash_fs_path(self) -> Path:
if self._hashfs_path is None:
self._hashfs_path = self.data_store_path / "hash_fs"
return self._hashfs_path
@property
def hashfs(self) -> HashFS:
if self._hashfs is None:
self._hashfs = HashFS(
self.hash_fs_path,
depth=DEFAULT_HASHFS_DEPTH,
width=DEFAULT_HASHFS_WIDTH,
algorithm=DEFAULT_HASH_FS_ALGORITHM,
)
return self._hashfs
def get_path(
self, entity_type: Optional[EntityType] = None, base_path: Optional[Path] = None
) -> Path:
if base_path is None:
if entity_type is None:
result = self.data_store_path
else:
result = self.data_store_path / entity_type.value
else:
if entity_type is None:
result = base_path
else:
result = base_path / entity_type.value
result.mkdir(parents=True, exist_ok=True)
return result
def _retrieve_environment_details(
self, env_type: str, env_hash: int
) -> Mapping[str, Any]:
base_path = self.get_path(entity_type=EntityType.ENVIRONMENT)
env_details_file = base_path / f"{env_type}_{env_hash}.json"
if not env_details_file.exists():
raise Exception(
f"Can't load environment details, file does not exist: {env_details_file.as_posix()}"
)
environment = orjson.loads(env_details_file.read_text())
return environment
def find_matching_job_record(
self, inputs_manifest: InputsManifest
) -> Optional[JobRecord]:
return self._retrieve_job_record(
manifest_hash=inputs_manifest.manifest_hash,
jobs_hash=inputs_manifest.job_hash,
)
def _retrieve_job_record(
self, manifest_hash: int, jobs_hash: int
) -> Optional[JobRecord]:
base_path = self.get_path(entity_type=EntityType.MANIFEST)
manifest_folder = base_path / str(manifest_hash)
if not manifest_folder.exists():
return None
manifest_file = manifest_folder / "manifest.json"
if not manifest_file.exists():
raise Exception(
f"No 'manifests.json' file for manifest with hash: {manifest_hash}"
)
manifest_data = orjson.loads(manifest_file.read_text())
job_folder = manifest_folder / str(jobs_hash)
if not job_folder.exists():
return None
inputs_file_name = job_folder / "inputs.json"
if not inputs_file_name.exists():
raise Exception(
f"No 'inputs.json' file for manifest/inputs hash-combo: {manifest_hash} / {jobs_hash}"
)
inputs_data = {
k: uuid.UUID(v)
for k, v in orjson.loads(inputs_file_name.read_text()).items()
}
outputs = {}
for output_file in job_folder.glob("output__*.json"):
full_output_name = output_file.name[8:]
start_value_id = full_output_name.find("__value_id__")
output_name = full_output_name[0:start_value_id]
value_id_str = full_output_name[start_value_id + 12 : -5] # noqa
value_id = uuid.UUID(value_id_str)
outputs[output_name] = value_id
job_id = ID_REGISTRY.generate(obj_type=JobRecord, desc="fake job id")
job_record = JobRecord(
job_id=job_id,
module_type=manifest_data["module_type"],
module_config=manifest_data["module_config"],
inputs=inputs_data,
outputs=outputs,
)
return job_record
def _find_values_with_hash(
self,
value_hash: int,
value_size: Optional[int] = None,
data_type_name: Optional[str] = None,
) -> Set[uuid.UUID]:
value_data_folder = self.get_path(entity_type=EntityType.VALUE_DATA)
glob = f"*/{value_hash}/value_id__*.json"
matches = list(value_data_folder.glob(glob))
result = set()
for match in matches:
if not match.is_symlink():
log_message(
f"Ignoring value_id file, not a symlink: {match.as_posix()}"
)
continue
uuid_str = match.name[10:-5]
value_id = uuid.UUID(uuid_str)
result.add(value_id)
return result
def _find_destinies_for_value(
self, value_id: uuid.UUID, alias_filter: Optional[str] = None
) -> Optional[Mapping[str, uuid.UUID]]:
destiny_dir = self.get_path(entity_type=EntityType.DESTINY_LINK)
destiny_value_dir = destiny_dir / str(value_id)
if not destiny_value_dir.exists():
return None
destinies = {}
for alias_link in destiny_value_dir.glob("*.json"):
assert alias_link.is_symlink()
alias = alias_link.name[0:-5]
resolved = alias_link.resolve()
value_id_str = resolved.parent.name
value_id = uuid.UUID(value_id_str)
destinies[alias] = value_id
return destinies
def _retrieve_all_value_ids(
self, data_type_name: Optional[str] = None
) -> Iterable[uuid.UUID]:
if data_type_name is not None:
raise NotImplementedError()
childs = self.get_path(entity_type=EntityType.VALUE).glob("*")
folders = [uuid.UUID(x.name) for x in childs if x.is_dir()]
return folders
def has_value(self, value_id: uuid.UUID) -> bool:
"""Check whether the specific value_id is persisted in this data store.
way to quickly determine whether a value id is valid for this data store.
Arguments:
value_id: the id of the value to check.
Returns:
whether this data store contains the value with the specified id
"""
base_path = (
self.get_path(entity_type=EntityType.VALUE)
/ str(value_id)
/ VALUE_DETAILS_FILE_NAME
)
return base_path.is_file()
def _retrieve_value_details(self, value_id: uuid.UUID) -> Mapping[str, Any]:
base_path = (
self.get_path(entity_type=EntityType.VALUE)
/ str(value_id)
/ VALUE_DETAILS_FILE_NAME
)
if not base_path.is_file():
raise Exception(
f"Can't retrieve details for value with id '{value_id}': no value with that id stored."
)
value_data = orjson.loads(base_path.read_text())
return value_data
def _retrieve_load_config(self, value: Value) -> LoadConfig:
base_path = self.get_path(entity_type=EntityType.VALUE_DATA)
data_dir = base_path / value.data_type_name / str(value.value_hash)
load_config_file = data_dir / ".load_config.json"
data = orjson.loads(load_config_file.read_text())
return LoadConfig(**data)
def retrieve_bytes(self, chunk_id: str, as_bytes: bool = True) -> Union[bytes, str]:
addr = self.hashfs.get(chunk_id)
if as_bytes:
return Path(addr.abspath).read_bytes()
else:
return addr.abspath
data_store_path: Path
property
readonly
¶hash_fs_path: Path
property
readonly
¶hashfs: HashFS
property
readonly
¶
_config_cls (ArchiveConfig)
private
pydantic-model
¶find_matching_job_record(self, inputs_manifest)
¶Source code in kiara/registries/data/data_store/filesystem_store.py
def find_matching_job_record(
self, inputs_manifest: InputsManifest
) -> Optional[JobRecord]:
return self._retrieve_job_record(
manifest_hash=inputs_manifest.manifest_hash,
jobs_hash=inputs_manifest.job_hash,
)
get_path(self, entity_type=None, base_path=None)
¶Source code in kiara/registries/data/data_store/filesystem_store.py
def get_path(
self, entity_type: Optional[EntityType] = None, base_path: Optional[Path] = None
) -> Path:
if base_path is None:
if entity_type is None:
result = self.data_store_path
else:
result = self.data_store_path / entity_type.value
else:
if entity_type is None:
result = base_path
else:
result = base_path / entity_type.value
result.mkdir(parents=True, exist_ok=True)
return result
has_value(self, value_id)
¶Check whether the specific value_id is persisted in this data store. way to quickly determine whether a value id is valid for this data store.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
value_id |
UUID |
the id of the value to check. |
required |
Returns:
| Type | Description |
|---|---|
bool |
whether this data store contains the value with the specified id |
Source code in kiara/registries/data/data_store/filesystem_store.py
def has_value(self, value_id: uuid.UUID) -> bool:
"""Check whether the specific value_id is persisted in this data store.
way to quickly determine whether a value id is valid for this data store.
Arguments:
value_id: the id of the value to check.
Returns:
whether this data store contains the value with the specified id
"""
base_path = (
self.get_path(entity_type=EntityType.VALUE)
/ str(value_id)
/ VALUE_DETAILS_FILE_NAME
)
return base_path.is_file()
is_writeable()
classmethod
¶Source code in kiara/registries/data/data_store/filesystem_store.py
@classmethod
def is_writeable(cls) -> bool:
return False
retrieve_bytes(self, chunk_id, as_bytes=True)
¶Source code in kiara/registries/data/data_store/filesystem_store.py
def retrieve_bytes(self, chunk_id: str, as_bytes: bool = True) -> Union[bytes, str]:
addr = self.hashfs.get(chunk_id)
if as_bytes:
return Path(addr.abspath).read_bytes()
else:
return addr.abspath
FilesystemDataStore (FileSystemDataArchive, BaseDataStore)
¶Data store that stores data as files on the local filesystem.
Source code in kiara/registries/data/data_store/filesystem_store.py
class FilesystemDataStore(FileSystemDataArchive, BaseDataStore):
"""Data store that stores data as files on the local filesystem."""
_archive_type_name = "filesystem_data_store"
def _persist_environment_details(
self, env_type: str, env_hash: int, env_data: Mapping[str, Any]
):
base_path = self.get_path(entity_type=EntityType.ENVIRONMENT)
env_details_file = base_path / f"{env_type}_{env_hash}.json"
if not env_details_file.exists():
env_details_file.write_text(orjson_dumps(env_data))
def _persist_load_config(self, value: Value, load_config: LoadConfig):
working_dir = self.get_path(entity_type=EntityType.VALUE_DATA)
data_dir = working_dir / value.data_type_name / str(value.value_hash)
load_config_file = data_dir / ".load_config.json"
data_dir.mkdir(exist_ok=True, parents=True)
load_config_file.write_text(load_config.json())
def _persist_value_details(self, value: Value):
value_dir = self.get_path(entity_type=EntityType.VALUE) / str(value.value_id)
if value_dir.exists():
raise Exception(
f"Can't persist value '{value.value_id}', value directory already exists: {value_dir}"
)
else:
value_dir.mkdir(parents=True, exist_ok=False)
value_file = value_dir / VALUE_DETAILS_FILE_NAME
value_data = value.dict()
value_file.write_text(orjson_dumps(value_data, option=orjson.OPT_NON_STR_KEYS))
def _persist_destiny_backlinks(self, value: Value):
destiny_dir = self.get_path(entity_type=EntityType.DESTINY_LINK)
for value_id, backlink in value.destiny_backlinks.items():
destiny_value_dir = destiny_dir / str(value_id)
destiny_value_dir.mkdir(parents=True, exist_ok=True)
destiny_file = destiny_value_dir / f"{backlink}.json"
assert not destiny_file.exists()
value_dir = self.get_path(entity_type=EntityType.VALUE) / str(
value.value_id
)
value_file = value_dir / VALUE_DETAILS_FILE_NAME
assert value_file.exists()
destiny_file.symlink_to(value_file)
def _persist_bytes(self, bytes_structure: BytesStructure) -> BytesAliasStructure:
bytes_alias_map: Dict[str, List[str]] = {}
for key, bytes_list in bytes_structure.chunk_map.items():
if is_debug():
assert not isinstance(bytes_list, (bytes, str))
for chunk in bytes_list:
if isinstance(chunk, str):
addr = self.hashfs.put(chunk)
elif isinstance(chunk, bytes):
_chunk = BytesIO(chunk)
addr = self.hashfs.put(_chunk)
else:
addr = self.hashfs.put(chunk)
# raise Exception(
# f"Can't persist chunk: invalid type '{type(chunk)}'"
# )
# we don't want anyone writing to a stored file
os.chmod(addr.abspath, S_IREAD | S_IRGRP | S_IROTH)
bytes_alias_map.setdefault(key, []).append(addr.id)
alias_structure = BytesAliasStructure.construct(
data_type=bytes_structure.data_type,
data_type_config=bytes_structure.data_type_config,
chunk_id_map=bytes_alias_map,
)
return alias_structure
def _persist_value_data(
self, value: Value
) -> Tuple[LoadConfig, Optional[BytesStructure]]:
persist_op_type = self.kiara_context.operation_registry.operation_types.get(
"persist_value", None
)
if persist_op_type is None:
raise Exception(
"Can't persist value, 'persist_value' operation type not available."
)
op_type: PersistValueOperationType = self._kiara.operation_registry.get_operation_type("persist_value") # type: ignore
op = op_type.get_operation_for_data_type(value.value_schema.type)
working_dir = self.get_path(entity_type=EntityType.VALUE_DATA)
data_dir = working_dir / value.data_type_name / str(value.value_hash)
result = op.run(
kiara=self.kiara_context,
inputs={
"value": value,
"persistence_config": {"temp_dir": data_dir.as_posix()},
},
)
load_config: LoadConfig = result.get_value_data("load_config")
bytes_structure: BytesStructure = result.get_value_data("bytes_structure")
return load_config, bytes_structure
def _persist_value_pedigree(self, value: Value):
manifest_hash = value.pedigree.manifest_hash
jobs_hash = value.pedigree.job_hash
base_path = self.get_path(entity_type=EntityType.MANIFEST)
manifest_folder = base_path / str(manifest_hash)
manifest_folder.mkdir(parents=True, exist_ok=True)
manifest_info_file = manifest_folder / "manifest.json"
if not manifest_info_file.exists():
manifest_info_file.write_text(value.pedigree.manifest_data_as_json())
job_folder = manifest_folder / str(jobs_hash)
job_folder.mkdir(parents=True, exist_ok=True)
inputs_details_file_name = job_folder / "inputs.json"
if not inputs_details_file_name.exists():
inputs_details_file_name.write_text(orjson_dumps(value.pedigree.inputs))
outputs_file_name = (
job_folder
/ f"output__{value.pedigree_output_name}__value_id__{value.value_id}.json"
)
if outputs_file_name.exists():
# if value.pedigree_output_name == "__void__":
# return
# else:
raise Exception(f"Can't write value '{value.value_id}': already exists.")
else:
outputs_file_name.touch()
value_data_dir = (
self.get_path(entity_type=EntityType.VALUE_DATA)
/ value.value_schema.type
/ str(value.value_hash)
)
target_file = value_data_dir / f"value_id__{value.value_id}.json"
target_file.symlink_to(outputs_file_name)
destinies
special
¶
Classes¶
DestinyArchive (ABC)
¶Source code in kiara/registries/destinies/__init__.py
class DestinyArchive(abc.ABC):
@abc.abstractmethod
def get_destiny_archive_id(self) -> uuid.UUID:
pass
@abc.abstractmethod
def get_all_value_ids(self) -> Set[uuid.UUID]:
"""Retrun a list of all value ids that have destinies stored in this archive."""
@abc.abstractmethod
def get_destiny_aliases_for_value(self, value_id: uuid.UUID) -> Optional[Set[str]]:
"""Retrieve all the destinies for the specified value within this archive.
In case this archive discovers its value destinies dynamically, this can return 'None'.
"""
@abc.abstractmethod
def get_destiny(self, value_id: uuid.UUID, destiny: str) -> Destiny:
pass
get_all_value_ids(self)
¶Retrun a list of all value ids that have destinies stored in this archive.
Source code in kiara/registries/destinies/__init__.py
@abc.abstractmethod
def get_all_value_ids(self) -> Set[uuid.UUID]:
"""Retrun a list of all value ids that have destinies stored in this archive."""
get_destiny(self, value_id, destiny)
¶Source code in kiara/registries/destinies/__init__.py
@abc.abstractmethod
def get_destiny(self, value_id: uuid.UUID, destiny: str) -> Destiny:
pass
get_destiny_aliases_for_value(self, value_id)
¶Retrieve all the destinies for the specified value within this archive.
In case this archive discovers its value destinies dynamically, this can return 'None'.
Source code in kiara/registries/destinies/__init__.py
@abc.abstractmethod
def get_destiny_aliases_for_value(self, value_id: uuid.UUID) -> Optional[Set[str]]:
"""Retrieve all the destinies for the specified value within this archive.
In case this archive discovers its value destinies dynamically, this can return 'None'.
"""
get_destiny_archive_id(self)
¶Source code in kiara/registries/destinies/__init__.py
@abc.abstractmethod
def get_destiny_archive_id(self) -> uuid.UUID:
pass
DestinyStore (DestinyArchive)
¶Source code in kiara/registries/destinies/__init__.py
class DestinyStore(DestinyArchive):
@abc.abstractmethod
def persist_destiny(self, destiny: Destiny):
pass
persist_destiny(self, destiny)
¶Source code in kiara/registries/destinies/__init__.py
@abc.abstractmethod
def persist_destiny(self, destiny: Destiny):
pass
Modules¶
filesystem_store
¶logger
¶
FileSystemDestinyArchive (DestinyArchive)
¶Source code in kiara/registries/destinies/filesystem_store.py
class FileSystemDestinyArchive(DestinyArchive):
@classmethod
def create_from_kiara_context(cls, kiara: "Kiara"):
base_path = Path(kiara.context_config.data_directory) / "destiny_store"
base_path.mkdir(parents=True, exist_ok=True)
result = cls(base_path=base_path, store_id=kiara.id)
ID_REGISTRY.update_metadata(
result.get_destiny_archive_id(), kiara_id=kiara.id, obj=result
)
return result
def __init__(self, base_path: Path, store_id: uuid.UUID):
if not base_path.is_dir():
raise Exception(
f"Can't create file system archive instance, base path does not exist or is not a folder: {base_path.as_posix()}."
)
self._store_id: uuid.UUID = store_id
self._base_path: Path = base_path
self._destinies_path: Path = self._base_path / "destinies"
self._value_id_path: Path = self._base_path / "value_ids"
@property
def destiny_store_path(self) -> Path:
return self._base_path
def get_destiny_archive_id(self) -> uuid.UUID:
return self._store_id
def _translate_destiny_id_to_path(self, destiny_id: uuid.UUID) -> Path:
tokens = str(destiny_id).split("-")
destiny_path = (
self._destinies_path.joinpath(*tokens[0:-1]) / f"{tokens[-1]}.json"
)
return destiny_path
def _translate_destinies_path_to_id(self, destinies_path: Path) -> uuid.UUID:
relative = destinies_path.relative_to(self._destinies_path).as_posix()[:-5]
destninies_id = "-".join(relative.split(os.path.sep))
return uuid.UUID(destninies_id)
def _translate_value_id(self, value_id: uuid.UUID, destiny_alias: str) -> Path:
tokens = str(value_id).split("-")
value_id_path = self._value_id_path.joinpath(*tokens)
full_path = value_id_path / f"{destiny_alias}.json"
return full_path
def _translate_value_id_path(self, value_path: Path) -> uuid.UUID:
relative = value_path.relative_to(self._value_id_path)
value_id_str = "-".join(relative.as_posix().split(os.path.sep))
return uuid.UUID(value_id_str)
def _translate_alias_path(self, alias_path: Path) -> Tuple[uuid.UUID, str]:
value_id = self._translate_value_id_path(alias_path.parent)
alias = alias_path.name[0:-5]
return value_id, alias
def get_all_value_ids(self) -> Set[uuid.UUID]:
all_root_folders = self._value_id_path.glob("*/*/*/*/*")
result = set()
for folder in all_root_folders:
if not folder.is_dir():
continue
value_id = self._translate_value_id_path(folder)
result.add(value_id)
return result
def get_destiny_aliases_for_value(self, value_id: uuid.UUID) -> Set[str]:
tokens = str(value_id).split("-")
value_id_path = self._value_id_path.joinpath(*tokens)
aliases = value_id_path.glob("*.json")
return set(a.name[0:-5] for a in aliases)
def get_destiny(self, value_id: uuid.UUID, destiny_alias: str) -> Destiny:
tokens = str(value_id).split("-")
value_id_path = self._value_id_path.joinpath(*tokens)
destiny_path = value_id_path / f"{destiny_alias}.json"
destiny_data = orjson.loads(destiny_path.read_text())
destiny = Destiny.construct(**destiny_data)
return destiny
destiny_store_path: Path
property
readonly
¶create_from_kiara_context(kiara)
classmethod
¶Source code in kiara/registries/destinies/filesystem_store.py
@classmethod
def create_from_kiara_context(cls, kiara: "Kiara"):
base_path = Path(kiara.context_config.data_directory) / "destiny_store"
base_path.mkdir(parents=True, exist_ok=True)
result = cls(base_path=base_path, store_id=kiara.id)
ID_REGISTRY.update_metadata(
result.get_destiny_archive_id(), kiara_id=kiara.id, obj=result
)
return result
get_all_value_ids(self)
¶Retrun a list of all value ids that have destinies stored in this archive.
Source code in kiara/registries/destinies/filesystem_store.py
def get_all_value_ids(self) -> Set[uuid.UUID]:
all_root_folders = self._value_id_path.glob("*/*/*/*/*")
result = set()
for folder in all_root_folders:
if not folder.is_dir():
continue
value_id = self._translate_value_id_path(folder)
result.add(value_id)
return result
get_destiny(self, value_id, destiny_alias)
¶Source code in kiara/registries/destinies/filesystem_store.py
def get_destiny(self, value_id: uuid.UUID, destiny_alias: str) -> Destiny:
tokens = str(value_id).split("-")
value_id_path = self._value_id_path.joinpath(*tokens)
destiny_path = value_id_path / f"{destiny_alias}.json"
destiny_data = orjson.loads(destiny_path.read_text())
destiny = Destiny.construct(**destiny_data)
return destiny
get_destiny_aliases_for_value(self, value_id)
¶Retrieve all the destinies for the specified value within this archive.
In case this archive discovers its value destinies dynamically, this can return 'None'.
Source code in kiara/registries/destinies/filesystem_store.py
def get_destiny_aliases_for_value(self, value_id: uuid.UUID) -> Set[str]:
tokens = str(value_id).split("-")
value_id_path = self._value_id_path.joinpath(*tokens)
aliases = value_id_path.glob("*.json")
return set(a.name[0:-5] for a in aliases)
get_destiny_archive_id(self)
¶Source code in kiara/registries/destinies/filesystem_store.py
def get_destiny_archive_id(self) -> uuid.UUID:
return self._store_id
FileSystemDestinyStore (FileSystemDestinyArchive, DestinyStore)
¶Source code in kiara/registries/destinies/filesystem_store.py
class FileSystemDestinyStore(FileSystemDestinyArchive, DestinyStore):
def persist_destiny(self, destiny: Destiny):
destiny_path = self._translate_destiny_id_to_path(destiny_id=destiny.destiny_id)
destiny_path.parent.mkdir(parents=True, exist_ok=True)
destiny_path.write_text(destiny.json())
for value_id in destiny.fixed_inputs.values():
path = self._translate_value_id(
value_id=value_id, destiny_alias=destiny.destiny_alias
)
if path.exists():
logger.debug("replace.destiny.file", path=path.as_posix())
path.unlink()
# raise Exception(
# f"Can't persist destiny '{destiny.destiny_id}': already persisted."
# )
path.parent.mkdir(parents=True, exist_ok=True)
path.symlink_to(destiny_path)
persist_destiny(self, destiny)
¶Source code in kiara/registries/destinies/filesystem_store.py
def persist_destiny(self, destiny: Destiny):
destiny_path = self._translate_destiny_id_to_path(destiny_id=destiny.destiny_id)
destiny_path.parent.mkdir(parents=True, exist_ok=True)
destiny_path.write_text(destiny.json())
for value_id in destiny.fixed_inputs.values():
path = self._translate_value_id(
value_id=value_id, destiny_alias=destiny.destiny_alias
)
if path.exists():
logger.debug("replace.destiny.file", path=path.as_posix())
path.unlink()
# raise Exception(
# f"Can't persist destiny '{destiny.destiny_id}': already persisted."
# )
path.parent.mkdir(parents=True, exist_ok=True)
path.symlink_to(destiny_path)
registry
¶
DestinyRegistry
¶Source code in kiara/registries/destinies/registry.py
class DestinyRegistry(object):
def __init__(self, kiara: "Kiara"):
self._kiara: Kiara = kiara
self._destiny_archives: Dict[str, DestinyArchive] = {}
self._default_destiny_store: Optional[str] = None
default_metadata_archive = FileSystemDestinyStore.create_from_kiara_context(
self._kiara
)
self.register_destiny_archive("metadata", default_metadata_archive)
self._all_values: Optional[Dict[uuid.UUID, Set[str]]] = None
self._cached_value_aliases: Dict[uuid.UUID, Dict[str, Optional[Destiny]]] = {}
self._destinies: Dict[uuid.UUID, Destiny] = {}
self._destinies_by_value: Dict[uuid.UUID, Dict[str, Destiny]] = {}
self._destiny_store_map: Dict[uuid.UUID, str] = {}
@property
def default_destiny_store(self) -> DestinyStore:
if self._default_destiny_store is None:
raise Exception("No default destiny store set (yet).")
return self._destiny_archives[self._default_destiny_store] # type: ignore
def register_destiny_archive(self, registered_name: str, alias_store: DestinyStore):
if not registered_name.isalnum():
raise Exception(
f"Can't register destiny archive with name '{registered_name}: name must only contain alphanumeric characters.'"
)
if registered_name in self._destiny_archives.keys():
raise Exception(
f"Can't register alias store, store id already registered: {registered_name}."
)
self._destiny_archives[registered_name] = alias_store
if self._default_destiny_store is None and isinstance(
alias_store, DestinyStore
):
self._default_destiny_store = registered_name
def _split_alias(self, alias: str) -> Tuple[str, str]:
if "." not in alias:
assert self._default_destiny_store is not None
return self._default_destiny_store, alias
store_id, rest = alias.split(".", maxsplit=1)
if store_id not in self._destiny_archives.keys():
raise Exception(
f"Invalid alias '{alias}', no store with prefix '{store_id}' registered. Available prefixes: {', '.join(self._destiny_archives.keys())}"
)
return (store_id, rest)
def add_destiny(
self,
destiny_alias: str,
values: Dict[str, uuid.UUID],
manifest: Manifest,
result_field_name: Optional[str] = None,
) -> Destiny:
"""Add a destiny for one (or in some rare cases several) values.
A destiny alias must be unique for every one of the involved input values.
"""
if not values:
raise Exception("Can't add destiny, no values provided.")
store_id, alias = self._split_alias(destiny_alias)
destiny = Destiny.create_from_values(
kiara=self._kiara,
destiny_alias=alias,
manifest=manifest,
result_field_name=result_field_name,
values=values,
)
for value_id in destiny.fixed_inputs.values():
self._destinies[destiny.destiny_id] = destiny
# TODO: store history?
self._destinies_by_value.setdefault(value_id, {})[destiny_alias] = destiny
self._cached_value_aliases.setdefault(value_id, {})[destiny_alias] = destiny
self._destiny_store_map[destiny.destiny_id] = store_id
return destiny
def get_destiny(self, value_id: uuid.UUID, destiny_alias: str) -> Destiny:
destiny = self._destinies_by_value.get(value_id, {}).get(destiny_alias, None)
if destiny is None:
raise Exception(
f"No destiny '{destiny_alias}' available for value '{value_id}'."
)
return destiny
@property
def _all_values_store_map(self) -> Dict[uuid.UUID, Set[str]]:
if self._all_values is not None:
return self._all_values
all_values: Dict[uuid.UUID, Set[str]] = {}
for archive_id, archive in self._destiny_archives.items():
all_value_ids = archive.get_all_value_ids()
for v_id in all_value_ids:
all_values.setdefault(v_id, set()).add(archive_id)
self._all_values = all_values
return self._all_values
@property
def all_values(self) -> Iterable[uuid.UUID]:
all_stored_values = set(self._all_values_store_map.keys())
all_stored_values.update(self._destinies_by_value.keys())
return all_stored_values
def get_destiny_aliases_for_value(
self, value_id: uuid.UUID, alias_filter: Optional[str] = None
) -> Iterable[str]:
# TODO: cache the result of this
if alias_filter is not None:
raise NotImplementedError()
all_stores = self._all_values_store_map.get(value_id)
aliases: Set[str] = set()
if all_stores:
for prefix in all_stores:
all_aliases = self._destiny_archives[
prefix
].get_destiny_aliases_for_value(value_id=value_id)
if all_aliases is not None:
aliases.update((f"{prefix}.{a}" for a in all_aliases))
current = self._destinies_by_value.get(value_id, None)
if current:
aliases.update(current.keys())
return sorted(aliases)
# def get_destinies_for_value(
# self,
# value_id: uuid.UUID,
# destiny_alias_filter: Optional[str] = None
# ) -> Mapping[str, Destiny]:
#
#
#
# return self._destinies_by_value.get(value_id, {})
def resolve_destiny(self, destiny: Destiny) -> Value:
results = self._kiara.job_registry.execute_and_retrieve(
manifest=destiny, inputs=destiny.merged_inputs
)
value = results.get_value_obj(field_name=destiny.result_field_name)
destiny.result_value_id = value.value_id
return value
def attach_as_property(
self,
destiny: Union[uuid.UUID, Destiny],
field_names: Optional[Iterable[str]] = None,
):
if field_names:
raise NotImplementedError()
if isinstance(destiny, uuid.UUID):
destiny = self._destinies[destiny]
values = self._kiara.data_registry.load_values(destiny.fixed_inputs)
already_stored: List[uuid.UUID] = []
for v in values.values():
if v.is_stored:
already_stored.append(v.value_id)
if already_stored:
stored = (str(v) for v in already_stored)
raise Exception(
f"Can't attach destiny as property, value(s) already stored: {', '.join(stored)}"
)
store_id = self._destiny_store_map[destiny.destiny_id]
full_path = f"{store_id}.{destiny.destiny_alias}"
for v in values.values():
assert destiny.result_value_id is not None
v.add_property(
value_id=destiny.result_value_id,
property_path=full_path,
add_origin_to_property_value=True,
)
def store_destiny(self, destiny_id: Union[Destiny, uuid.UUID]):
try:
_destiny_id: uuid.UUID = destiny_id.destiny_id # type: ignore
except Exception:
# just in case this is a 'Destiny' object
_destiny_id = destiny_id # type: ignore
store_id = self._destiny_store_map[_destiny_id]
destiny = self._destinies[_destiny_id]
store: DestinyStore = self._destiny_archives[store_id] # type: ignore
if not isinstance(store, DestinyStore):
full_alias = f"{store_id}.{destiny.destiny_alias}"
raise Exception(
f"Can't store destiny '{full_alias}': prefix '{store_id}' not writable in this kiara context."
)
store.persist_destiny(destiny=destiny)
all_values: Iterable[uuid.UUID]
property
readonly
¶default_destiny_store: DestinyStore
property
readonly
¶add_destiny(self, destiny_alias, values, manifest, result_field_name=None)
¶Add a destiny for one (or in some rare cases several) values.
A destiny alias must be unique for every one of the involved input values.
Source code in kiara/registries/destinies/registry.py
def add_destiny(
self,
destiny_alias: str,
values: Dict[str, uuid.UUID],
manifest: Manifest,
result_field_name: Optional[str] = None,
) -> Destiny:
"""Add a destiny for one (or in some rare cases several) values.
A destiny alias must be unique for every one of the involved input values.
"""
if not values:
raise Exception("Can't add destiny, no values provided.")
store_id, alias = self._split_alias(destiny_alias)
destiny = Destiny.create_from_values(
kiara=self._kiara,
destiny_alias=alias,
manifest=manifest,
result_field_name=result_field_name,
values=values,
)
for value_id in destiny.fixed_inputs.values():
self._destinies[destiny.destiny_id] = destiny
# TODO: store history?
self._destinies_by_value.setdefault(value_id, {})[destiny_alias] = destiny
self._cached_value_aliases.setdefault(value_id, {})[destiny_alias] = destiny
self._destiny_store_map[destiny.destiny_id] = store_id
return destiny
attach_as_property(self, destiny, field_names=None)
¶Source code in kiara/registries/destinies/registry.py
def attach_as_property(
self,
destiny: Union[uuid.UUID, Destiny],
field_names: Optional[Iterable[str]] = None,
):
if field_names:
raise NotImplementedError()
if isinstance(destiny, uuid.UUID):
destiny = self._destinies[destiny]
values = self._kiara.data_registry.load_values(destiny.fixed_inputs)
already_stored: List[uuid.UUID] = []
for v in values.values():
if v.is_stored:
already_stored.append(v.value_id)
if already_stored:
stored = (str(v) for v in already_stored)
raise Exception(
f"Can't attach destiny as property, value(s) already stored: {', '.join(stored)}"
)
store_id = self._destiny_store_map[destiny.destiny_id]
full_path = f"{store_id}.{destiny.destiny_alias}"
for v in values.values():
assert destiny.result_value_id is not None
v.add_property(
value_id=destiny.result_value_id,
property_path=full_path,
add_origin_to_property_value=True,
)
get_destiny(self, value_id, destiny_alias)
¶Source code in kiara/registries/destinies/registry.py
def get_destiny(self, value_id: uuid.UUID, destiny_alias: str) -> Destiny:
destiny = self._destinies_by_value.get(value_id, {}).get(destiny_alias, None)
if destiny is None:
raise Exception(
f"No destiny '{destiny_alias}' available for value '{value_id}'."
)
return destiny
get_destiny_aliases_for_value(self, value_id, alias_filter=None)
¶Source code in kiara/registries/destinies/registry.py
def get_destiny_aliases_for_value(
self, value_id: uuid.UUID, alias_filter: Optional[str] = None
) -> Iterable[str]:
# TODO: cache the result of this
if alias_filter is not None:
raise NotImplementedError()
all_stores = self._all_values_store_map.get(value_id)
aliases: Set[str] = set()
if all_stores:
for prefix in all_stores:
all_aliases = self._destiny_archives[
prefix
].get_destiny_aliases_for_value(value_id=value_id)
if all_aliases is not None:
aliases.update((f"{prefix}.{a}" for a in all_aliases))
current = self._destinies_by_value.get(value_id, None)
if current:
aliases.update(current.keys())
return sorted(aliases)
register_destiny_archive(self, registered_name, alias_store)
¶Source code in kiara/registries/destinies/registry.py
def register_destiny_archive(self, registered_name: str, alias_store: DestinyStore):
if not registered_name.isalnum():
raise Exception(
f"Can't register destiny archive with name '{registered_name}: name must only contain alphanumeric characters.'"
)
if registered_name in self._destiny_archives.keys():
raise Exception(
f"Can't register alias store, store id already registered: {registered_name}."
)
self._destiny_archives[registered_name] = alias_store
if self._default_destiny_store is None and isinstance(
alias_store, DestinyStore
):
self._default_destiny_store = registered_name
resolve_destiny(self, destiny)
¶Source code in kiara/registries/destinies/registry.py
def resolve_destiny(self, destiny: Destiny) -> Value:
results = self._kiara.job_registry.execute_and_retrieve(
manifest=destiny, inputs=destiny.merged_inputs
)
value = results.get_value_obj(field_name=destiny.result_field_name)
destiny.result_value_id = value.value_id
return value
store_destiny(self, destiny_id)
¶Source code in kiara/registries/destinies/registry.py
def store_destiny(self, destiny_id: Union[Destiny, uuid.UUID]):
try:
_destiny_id: uuid.UUID = destiny_id.destiny_id # type: ignore
except Exception:
# just in case this is a 'Destiny' object
_destiny_id = destiny_id # type: ignore
store_id = self._destiny_store_map[_destiny_id]
destiny = self._destinies[_destiny_id]
store: DestinyStore = self._destiny_archives[store_id] # type: ignore
if not isinstance(store, DestinyStore):
full_alias = f"{store_id}.{destiny.destiny_alias}"
raise Exception(
f"Can't store destiny '{full_alias}': prefix '{store_id}' not writable in this kiara context."
)
store.persist_destiny(destiny=destiny)
environment
special
¶
Classes¶
EnvironmentRegistry
¶Source code in kiara/registries/environment/__init__.py
class EnvironmentRegistry(object):
_instance = None
@classmethod
def instance(cls):
"""The default *kiara* context. In most cases, it's recommended you create and manage your own, though."""
if cls._instance is None:
cls._instance = EnvironmentRegistry()
return cls._instance
def __init__(
self,
):
self._environments: Optional[Dict[str, RuntimeEnvironment]] = None
self._full_env_model: Optional[BaseModel] = None
def get_environment_for_hash(self, env_hash: int) -> RuntimeEnvironment:
envs = [
env for env in self.environments.values() if env.model_data_hash == env_hash
]
if len(envs) == 0:
raise Exception(f"No environment with hash '{env_hash}' available.")
elif len(envs) > 1:
raise Exception(
f"Multipe environments with hash '{env_hash}' available. This is most likely a bug."
)
return envs[0]
@property
def environments(self) -> Mapping[str, RuntimeEnvironment]:
"""Return all environments in this kiara runtime context."""
if self._environments is not None:
return self._environments
import kiara.models.runtime_environment.kiara # noqa
import kiara.models.runtime_environment.operating_system # noqa
import kiara.models.runtime_environment.python # noqa
subclasses: Iterable[Type[RuntimeEnvironment]] = _get_all_subclasses(
RuntimeEnvironment # type: ignore
)
envs = {}
for sc in subclasses:
if inspect.isabstract(sc):
if is_debug():
logger.warning("class_loading.ignore_subclass", subclass=sc)
else:
logger.debug("class_loading.ignore_subclass", subclass=sc)
name = sc.get_environment_type_name()
envs[name] = sc.create_environment_model()
self._environments = envs
return self._environments
@property
def full_model(self) -> BaseModel:
"""A model containing all environment data, incl. schemas and hashes of each sub-environment."""
if self._full_env_model is not None:
return self._full_env_model
attrs = {k: (v.__class__, ...) for k, v in self.environments.items()}
models = {}
hashes = {}
schemas = {}
for k, v in attrs.items():
name = to_camel_case(f"{k}_environment")
k_cls: Type[RuntimeEnvironment] = create_model(
name,
__base__=v[0],
metadata_hash=(
str,
Field(
description="The hash for this metadata (excl. this and the 'metadata_schema' field)."
),
),
metadata_schema=(
str,
Field(
description="JsonSchema describing this metadata (excl. this and the 'metadata_hash' field)."
),
),
)
models[k] = (
k_cls,
Field(description=f"Metadata describing the {k} environment."),
)
schemas[k] = v[0].schema_json()
hashes[k] = self.environments[k].model_data_hash
cls: Type[BaseModel] = create_model("KiaraRuntimeInfo", **models) # type: ignore
data = {}
for k2, v2 in self.environments.items():
d = v2.dict()
assert "metadata_hash" not in d.keys()
assert "metadata_schema" not in d.keys()
d["metadata_hash"] = hashes[k2]
d["metadata_schema"] = schemas[k]
data[k2] = d
model = cls.construct(**data) # type: ignore
self._full_env_model = model
return self._full_env_model
def create_renderable(self, **config: Any):
full_details = config.get("full_details", False)
table = Table(show_header=True, box=box.SIMPLE)
table.add_column("environment key", style="b")
table.add_column("details")
for env_name, env in self.environments.items():
renderable = env.create_renderable(summary=not full_details)
table.add_row(env_name, renderable)
return table
environments: Mapping[str, kiara.models.runtime_environment.RuntimeEnvironment]
property
readonly
¶Return all environments in this kiara runtime context.
full_model: BaseModel
property
readonly
¶A model containing all environment data, incl. schemas and hashes of each sub-environment.
create_renderable(self, **config)
¶Source code in kiara/registries/environment/__init__.py
def create_renderable(self, **config: Any):
full_details = config.get("full_details", False)
table = Table(show_header=True, box=box.SIMPLE)
table.add_column("environment key", style="b")
table.add_column("details")
for env_name, env in self.environments.items():
renderable = env.create_renderable(summary=not full_details)
table.add_row(env_name, renderable)
return table
get_environment_for_hash(self, env_hash)
¶Source code in kiara/registries/environment/__init__.py
def get_environment_for_hash(self, env_hash: int) -> RuntimeEnvironment:
envs = [
env for env in self.environments.values() if env.model_data_hash == env_hash
]
if len(envs) == 0:
raise Exception(f"No environment with hash '{env_hash}' available.")
elif len(envs) > 1:
raise Exception(
f"Multipe environments with hash '{env_hash}' available. This is most likely a bug."
)
return envs[0]
instance()
classmethod
¶The default kiara context. In most cases, it's recommended you create and manage your own, though.
Source code in kiara/registries/environment/__init__.py
@classmethod
def instance(cls):
"""The default *kiara* context. In most cases, it's recommended you create and manage your own, though."""
if cls._instance is None:
cls._instance = EnvironmentRegistry()
return cls._instance
events
special
¶
AsyncEventListener (Protocol)
¶Source code in kiara/registries/events/__init__.py
class AsyncEventListener(Protocol):
def wait_for_processing(self, processing_id: Any):
pass
wait_for_processing(self, processing_id)
¶Source code in kiara/registries/events/__init__.py
def wait_for_processing(self, processing_id: Any):
pass
EventListener (Protocol)
¶Source code in kiara/registries/events/__init__.py
class EventListener(Protocol):
def handle_events(self, *events: KiaraEvent) -> Any:
pass
handle_events(self, *events)
¶Source code in kiara/registries/events/__init__.py
def handle_events(self, *events: KiaraEvent) -> Any:
pass
EventProducer (Protocol)
¶Source code in kiara/registries/events/__init__.py
class EventProducer(Protocol):
pass
# def suppoerted_event_types(self) -> Iterable[Type[KiaraEvent]]:
# pass
metadata
¶
CreateMetadataDestinies
¶Source code in kiara/registries/events/metadata.py
class CreateMetadataDestinies(object):
def __init__(self, kiara: "Kiara"):
self._kiara: Kiara = kiara
self._skip_internal_types: bool = True
def supported_event_types(self) -> Iterable[str]:
return ["value_created", "value_pre_store"]
def handle_events(self, *events: KiaraEvent) -> Any:
for event in events:
if event.get_event_type() == "value_created": # type: ignore
self.attach_metadata(event.value) # type: ignore
for event in events:
if event.get_event_type() == "value_pre_store": # type: ignore
self.resolve_all_metadata(event.value) # type: ignore
def attach_metadata(self, value: Value):
assert not value.is_stored
if self._skip_internal_types:
lineage = self._kiara.type_registry.get_type_lineage(
value.value_schema.type
)
if "any" not in lineage:
return
op_type: ExtractMetadataOperationType = self._kiara.operation_registry.get_operation_type("extract_metadata") # type: ignore
operations = op_type.get_operations_for_data_type(value.value_schema.type)
for metadata_key, op in operations.items():
op_details: ExtractMetadataDetails = op.operation_details # type: ignore
input_field_name = op_details.input_field_name
result_field_name = op_details.result_field_name
self._kiara.destiny_registry.add_destiny(
destiny_alias=f"metadata.{metadata_key}",
values={input_field_name: value.value_id},
manifest=op,
result_field_name=result_field_name,
)
def resolve_all_metadata(self, value: Value):
assert not value.is_stored
aliases = self._kiara.destiny_registry.get_destiny_aliases_for_value(
value_id=value.value_id
)
for alias in aliases:
destiny = self._kiara.destiny_registry.get_destiny(
value_id=value.value_id, destiny_alias=alias
)
self._kiara.destiny_registry.resolve_destiny(destiny)
self._kiara.destiny_registry.attach_as_property(destiny)
attach_metadata(self, value)
¶Source code in kiara/registries/events/metadata.py
def attach_metadata(self, value: Value):
assert not value.is_stored
if self._skip_internal_types:
lineage = self._kiara.type_registry.get_type_lineage(
value.value_schema.type
)
if "any" not in lineage:
return
op_type: ExtractMetadataOperationType = self._kiara.operation_registry.get_operation_type("extract_metadata") # type: ignore
operations = op_type.get_operations_for_data_type(value.value_schema.type)
for metadata_key, op in operations.items():
op_details: ExtractMetadataDetails = op.operation_details # type: ignore
input_field_name = op_details.input_field_name
result_field_name = op_details.result_field_name
self._kiara.destiny_registry.add_destiny(
destiny_alias=f"metadata.{metadata_key}",
values={input_field_name: value.value_id},
manifest=op,
result_field_name=result_field_name,
)
handle_events(self, *events)
¶Source code in kiara/registries/events/metadata.py
def handle_events(self, *events: KiaraEvent) -> Any:
for event in events:
if event.get_event_type() == "value_created": # type: ignore
self.attach_metadata(event.value) # type: ignore
for event in events:
if event.get_event_type() == "value_pre_store": # type: ignore
self.resolve_all_metadata(event.value) # type: ignore
resolve_all_metadata(self, value)
¶Source code in kiara/registries/events/metadata.py
def resolve_all_metadata(self, value: Value):
assert not value.is_stored
aliases = self._kiara.destiny_registry.get_destiny_aliases_for_value(
value_id=value.value_id
)
for alias in aliases:
destiny = self._kiara.destiny_registry.get_destiny(
value_id=value.value_id, destiny_alias=alias
)
self._kiara.destiny_registry.resolve_destiny(destiny)
self._kiara.destiny_registry.attach_as_property(destiny)
supported_event_types(self)
¶Source code in kiara/registries/events/metadata.py
def supported_event_types(self) -> Iterable[str]:
return ["value_created", "value_pre_store"]
registry
¶
AllEvents (KiaraEvent)
pydantic-model
¶Source code in kiara/registries/events/registry.py
class AllEvents(KiaraEvent):
pass
EventRegistry
¶Source code in kiara/registries/events/registry.py
class EventRegistry(object):
def __init__(self, kiara: "Kiara"):
self._kiara: Kiara = kiara
self._producers: Dict[uuid.UUID, EventProducer] = {}
self._listeners: Dict[uuid.UUID, EventListener] = {}
self._subscriptions: Dict[uuid.UUID, List[str]] = {}
def add_producer(self, producer: EventProducer) -> Callable:
producer_id = ID_REGISTRY.generate(
obj=producer, comment="adding event producer"
)
func = partial(self.handle_events, producer_id)
return func
def add_listener(self, listener, *subscriptions: str):
if not subscriptions:
_subscriptions = ["*"]
else:
_subscriptions = list(subscriptions)
listener_id = ID_REGISTRY.generate(
obj=listener, comment="adding event listener"
)
self._listeners[listener_id] = listener
self._subscriptions[listener_id] = _subscriptions
def _matches_subscription(
self, events: Iterable[KiaraEvent], subscriptions: Iterable[str]
) -> Iterable[KiaraEvent]:
result = []
for subscription in subscriptions:
for event in events:
match = fnmatch.filter([event.get_event_type()], subscription)
if match:
result.append(event)
return result
def handle_events(self, producer_id: uuid.UUID, *events: KiaraEvent):
event_targets: Dict[uuid.UUID, List[KiaraEvent]] = {}
for l_id, listener in self._listeners.items():
matches = self._matches_subscription(
events=events, subscriptions=self._subscriptions[l_id]
)
if matches:
event_targets.setdefault(l_id, []).extend(matches)
responses = {}
for l_id, l_events in event_targets.items():
listener = self._listeners[l_id]
response = listener.handle_events(*l_events)
responses[l_id] = response
for l_id, response in responses.items():
if response is None:
continue
a_listener: AsyncEventListener = self._listeners[l_id] # type: ignore
if not hasattr(a_listener, "wait_for_processing"):
raise Exception(
"Can't wait for processing of event for listener: listener does not provide 'wait_for_processing' method."
)
a_listener.wait_for_processing(response)
add_listener(self, listener, *subscriptions)
¶Source code in kiara/registries/events/registry.py
def add_listener(self, listener, *subscriptions: str):
if not subscriptions:
_subscriptions = ["*"]
else:
_subscriptions = list(subscriptions)
listener_id = ID_REGISTRY.generate(
obj=listener, comment="adding event listener"
)
self._listeners[listener_id] = listener
self._subscriptions[listener_id] = _subscriptions
add_producer(self, producer)
¶Source code in kiara/registries/events/registry.py
def add_producer(self, producer: EventProducer) -> Callable:
producer_id = ID_REGISTRY.generate(
obj=producer, comment="adding event producer"
)
func = partial(self.handle_events, producer_id)
return func
handle_events(self, producer_id, *events)
¶Source code in kiara/registries/events/registry.py
def handle_events(self, producer_id: uuid.UUID, *events: KiaraEvent):
event_targets: Dict[uuid.UUID, List[KiaraEvent]] = {}
for l_id, listener in self._listeners.items():
matches = self._matches_subscription(
events=events, subscriptions=self._subscriptions[l_id]
)
if matches:
event_targets.setdefault(l_id, []).extend(matches)
responses = {}
for l_id, l_events in event_targets.items():
listener = self._listeners[l_id]
response = listener.handle_events(*l_events)
responses[l_id] = response
for l_id, response in responses.items():
if response is None:
continue
a_listener: AsyncEventListener = self._listeners[l_id] # type: ignore
if not hasattr(a_listener, "wait_for_processing"):
raise Exception(
"Can't wait for processing of event for listener: listener does not provide 'wait_for_processing' method."
)
a_listener.wait_for_processing(response)
ids
special
¶
ID_REGISTRY
¶logger
¶
IdRegistry
¶Source code in kiara/registries/ids/__init__.py
class IdRegistry(object):
def __init__(self):
self._ids: Dict[uuid.UUID, Dict[Type, Dict[str, Any]]] = {}
self._objs: Dict[uuid.UUID, WeakValueDictionary[Type, Any]] = {}
def generate(
self,
id: Optional[uuid.UUID] = None,
obj_type: Optional[Type] = None,
obj: Optional[Any] = None,
**metadata: Any
):
if id is None:
id = uuid.uuid4()
if is_debug() or is_develop():
# logger.debug("generate.id", id=id, metadata=metadata)
if obj_type is None:
if obj:
obj_type = obj.__class__
else:
obj_type = NO_TYPE_MARKER
self._ids.setdefault(id, {}).setdefault(obj_type, {}).update(metadata)
if obj:
self._objs.setdefault(id, WeakValueDictionary())[obj_type] = obj
return id
def update_metadata(
self,
id: uuid.UUID,
obj_type: Optional[Type] = None,
obj: Optional[Any] = None,
**metadata
):
if not is_debug() and not is_develop():
return
if obj_type is None:
if obj:
obj_type = obj.__class__
else:
obj_type = NO_TYPE_MARKER
self._ids.setdefault(id, {}).setdefault(obj_type, {}).update(metadata)
if obj:
self._objs.setdefault(id, WeakValueDictionary())[obj_type] = obj
generate(self, id=None, obj_type=None, obj=None, **metadata)
¶Source code in kiara/registries/ids/__init__.py
def generate(
self,
id: Optional[uuid.UUID] = None,
obj_type: Optional[Type] = None,
obj: Optional[Any] = None,
**metadata: Any
):
if id is None:
id = uuid.uuid4()
if is_debug() or is_develop():
# logger.debug("generate.id", id=id, metadata=metadata)
if obj_type is None:
if obj:
obj_type = obj.__class__
else:
obj_type = NO_TYPE_MARKER
self._ids.setdefault(id, {}).setdefault(obj_type, {}).update(metadata)
if obj:
self._objs.setdefault(id, WeakValueDictionary())[obj_type] = obj
return id
update_metadata(self, id, obj_type=None, obj=None, **metadata)
¶Source code in kiara/registries/ids/__init__.py
def update_metadata(
self,
id: uuid.UUID,
obj_type: Optional[Type] = None,
obj: Optional[Any] = None,
**metadata
):
if not is_debug() and not is_develop():
return
if obj_type is None:
if obj:
obj_type = obj.__class__
else:
obj_type = NO_TYPE_MARKER
self._ids.setdefault(id, {}).setdefault(obj_type, {}).update(metadata)
if obj:
self._objs.setdefault(id, WeakValueDictionary())[obj_type] = obj
NO_TYPE_MARKER
¶Source code in kiara/registries/ids/__init__.py
class NO_TYPE_MARKER(object):
pass
jobs
special
¶
MANIFEST_SUB_PATH
¶logger
¶Classes¶
JobArchive (BaseArchive)
¶Source code in kiara/registries/jobs/__init__.py
class JobArchive(BaseArchive):
@abc.abstractmethod
def find_matching_job_record(
self, inputs_manifest: InputsManifest
) -> Optional[JobRecord]:
pass
find_matching_job_record(self, inputs_manifest)
¶Source code in kiara/registries/jobs/__init__.py
@abc.abstractmethod
def find_matching_job_record(
self, inputs_manifest: InputsManifest
) -> Optional[JobRecord]:
pass
JobRegistry
¶Source code in kiara/registries/jobs/__init__.py
class JobRegistry(object):
def __init__(self, kiara: "Kiara"):
self._kiara: Kiara = kiara
self._active_jobs: bidict[int, uuid.UUID] = bidict()
self._failed_jobs: Dict[int, uuid.UUID] = {}
self._finished_jobs: Dict[int, uuid.UUID] = {}
self._archived_records: Dict[uuid.UUID, JobRecord] = {}
self._processor: ModuleProcessor = SynchronousProcessor(kiara=self._kiara)
self._processor.register_job_status_listener(self)
self._job_archives: Dict[str, JobArchive] = {}
self._default_job_store: Optional[str] = None
self._event_callback = self._kiara.event_registry.add_producer(self)
# default_archive = FileSystemJobStore.create_from_kiara_context(self._kiara)
# self.register_job_archive(default_archive, store_alias=DEFAULT_STORE_MARKER)
# default_file_store = self._kiara.data_registry.get_archive(DEFAULT_STORE_MARKER)
# self.register_job_archive(default_file_store, store_alias="default_data_store") # type: ignore
def suppoerted_event_types(self) -> Iterable[Type[KiaraEvent]]:
return [JobArchiveAddedEvent, JobRecordPreStoreEvent, JobRecordStoredEvent]
def register_job_archive(self, archive: JobArchive, alias: Optional[str] = None):
if alias is None:
alias = str(archive.archive_id)
if alias in self._job_archives.keys():
raise Exception(
f"Can't register job store, store id already registered: {alias}."
)
self._job_archives[alias] = archive
is_store = False
is_default_store = False
if isinstance(archive, JobStore):
is_store = True
if self._default_job_store is None:
self._default_job_store = alias
event = JobArchiveAddedEvent.construct(
kiara_id=self._kiara.id,
job_archive_id=archive.archive_id,
job_archive_alias=alias,
is_store=is_store,
is_default_store=is_default_store,
)
self._event_callback(event)
@property
def default_job_store(self) -> str:
if self._default_job_store is None:
raise Exception("No default job store set (yet).")
return self._default_job_store # type: ignore
def get_archive(self, store_id: Optional[str] = None) -> JobArchive:
if store_id is None:
store_id = self.default_job_store
if store_id is None:
raise Exception("Can't retrieve deafult job archive, none set (yet).")
return self._job_archives[store_id]
def job_status_changed(
self, job_id: uuid.UUID, old_status: Optional[JobStatus], new_status: JobStatus
):
# print(f"JOB STATUS CHANGED: {job_id} - {old_status} - {new_status.value}")
if job_id in self._active_jobs.values() and new_status is JobStatus.FAILED:
job_hash = self._active_jobs.inverse.pop(job_id)
self._failed_jobs[job_hash] = job_id
elif job_id in self._active_jobs.values() and new_status is JobStatus.SUCCESS:
job_hash = self._active_jobs.inverse.pop(job_id)
job_record = self._processor.get_job_record(job_id)
self._finished_jobs[job_hash] = job_id
self._archived_records[job_id] = job_record
def store_job_record(self, job_id: uuid.UUID):
if job_id not in self._archived_records.keys():
raise Exception(
f"Can't store job with id '{job_id}': no job record with that id exists."
)
job_record = self._archived_records[job_id]
if job_record._is_stored:
logger.debug(
"ignore.store.job_record", reason="already stored", job_id=str(job_id)
)
return
logger.debug(
"store.job_record",
job_hash=job_record.job_hash,
module_type=job_record.module_type,
)
store: JobStore = self.get_archive() # type: ignore
if not isinstance(store, JobStore):
raise Exception("Can't store job record to archive: not writable.")
pre_store_event = JobRecordPreStoreEvent.construct(
kiara_id=self._kiara.id, job_record=job_record
)
self._event_callback(pre_store_event)
store.store_job_record(job_record)
stored_event = JobRecordStoredEvent.construct(
kiara_id=self._kiara.id, job_record=job_record
)
self._event_callback(stored_event)
def get_job_record_in_session(self, job_id: uuid.UUID) -> JobRecord:
return self._processor.get_job_record(job_id)
def find_matching_job_record(
self, inputs_manifest: InputsManifest
) -> Optional[uuid.UUID]:
"""Check if a job with same inputs manifest already ran some time before.
Arguments:
inputs_manifest: the manifest incl. inputs
Returns:
'None' if no such job exists, a (uuid) job-id if the job is currently running or has run in the past
"""
if inputs_manifest.job_hash in self._active_jobs.keys():
logger.debug("job.use_running")
return self._active_jobs[inputs_manifest.job_hash]
if inputs_manifest.job_hash in self._finished_jobs.keys():
job_id = self._finished_jobs[inputs_manifest.job_hash]
return job_id
matches = []
for store_id, archive in self._job_archives.items():
match = archive.find_matching_job_record(inputs_manifest=inputs_manifest)
if match:
matches.append(match)
if len(matches) == 0:
return None
elif len(matches) > 1:
raise Exception(
f"Multiple stores have a record for inputs manifest '{inputs_manifest}', this is not supported (yet)."
)
job_record = matches[0]
self._finished_jobs[inputs_manifest.job_hash] = job_record.job_id
self._archived_records[job_record.job_id] = job_record
logger.debug("job.use_cached")
return job_record.job_id
def prepare_job_config(
self, manifest: Manifest, inputs: Mapping[str, Any]
) -> JobConfig:
module = self._kiara.create_module(manifest=manifest)
job_config = JobConfig.create_from_module(
data_registry=self._kiara.data_registry, module=module, inputs=inputs
)
return job_config
def execute(
self, manifest: Manifest, inputs: Mapping[str, Any], wait: bool = False
) -> uuid.UUID:
job_config = self.prepare_job_config(manifest=manifest, inputs=inputs)
return self.execute_job(job_config, wait=wait)
def execute_job(self, job_config: JobConfig, wait: bool = False) -> uuid.UUID:
log = logger.bind(
module_type=job_config.module_type,
inputs={k: str(v) for k, v in job_config.inputs.items()},
job_hash=job_config.model_data_hash,
)
stored_job = self.find_matching_job_record(inputs_manifest=job_config)
if stored_job is not None:
return stored_job
log.debug("job.execute", inputs=job_config.inputs)
job_id = self._processor.create_job(job_config=job_config)
self._active_jobs[job_config.job_hash] = job_id
try:
self._processor.queue_job(job_id=job_id)
except Exception as e:
log.error("error.queue_job", job_id=job_id)
raise e
if wait:
self._processor.wait_for(job_id)
return job_id
def get_active_job(self, job_id: uuid.UUID) -> ActiveJob:
if job_id in self._active_jobs.keys() or job_id in self._failed_jobs.keys():
return self._processor.get_job(job_id)
else:
if job_id in self._archived_records.keys():
raise Exception(
f"Can't retrieve active job with id '{job_id}': job is archived."
)
else:
raise Exception(f"Can't retrieve job with id '{job_id}': no such job.")
def get_job_status(self, job_id: uuid.UUID) -> JobStatus:
if job_id in self._archived_records.keys():
return JobStatus.SUCCESS
elif job_id in self._failed_jobs.values():
return JobStatus.FAILED
return self._processor.get_job_status(job_id=job_id)
def wait_for(self, *job_id: uuid.UUID):
not_finished = (j for j in job_id if j not in self._archived_records.keys())
if not_finished:
self._processor.wait_for(*not_finished)
def retrieve_result(self, job_id: uuid.UUID) -> ValueMap:
if job_id not in self._archived_records.keys():
self._processor.wait_for(job_id)
job_record = self._archived_records[job_id]
results = self._kiara.data_registry.load_values(job_record.outputs)
return results
def execute_and_retrieve(
self, manifest: Manifest, inputs: Mapping[str, Any]
) -> ValueMap:
job_id = self.execute(manifest=manifest, inputs=inputs, wait=True)
results = self.retrieve_result(job_id=job_id)
return results
default_job_store: str
property
readonly
¶execute(self, manifest, inputs, wait=False)
¶Source code in kiara/registries/jobs/__init__.py
def execute(
self, manifest: Manifest, inputs: Mapping[str, Any], wait: bool = False
) -> uuid.UUID:
job_config = self.prepare_job_config(manifest=manifest, inputs=inputs)
return self.execute_job(job_config, wait=wait)
execute_and_retrieve(self, manifest, inputs)
¶Source code in kiara/registries/jobs/__init__.py
def execute_and_retrieve(
self, manifest: Manifest, inputs: Mapping[str, Any]
) -> ValueMap:
job_id = self.execute(manifest=manifest, inputs=inputs, wait=True)
results = self.retrieve_result(job_id=job_id)
return results
execute_job(self, job_config, wait=False)
¶Source code in kiara/registries/jobs/__init__.py
def execute_job(self, job_config: JobConfig, wait: bool = False) -> uuid.UUID:
log = logger.bind(
module_type=job_config.module_type,
inputs={k: str(v) for k, v in job_config.inputs.items()},
job_hash=job_config.model_data_hash,
)
stored_job = self.find_matching_job_record(inputs_manifest=job_config)
if stored_job is not None:
return stored_job
log.debug("job.execute", inputs=job_config.inputs)
job_id = self._processor.create_job(job_config=job_config)
self._active_jobs[job_config.job_hash] = job_id
try:
self._processor.queue_job(job_id=job_id)
except Exception as e:
log.error("error.queue_job", job_id=job_id)
raise e
if wait:
self._processor.wait_for(job_id)
return job_id
find_matching_job_record(self, inputs_manifest)
¶Check if a job with same inputs manifest already ran some time before.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
inputs_manifest |
InputsManifest |
the manifest incl. inputs |
required |
Returns:
| Type | Description |
|---|---|
Optional[uuid.UUID] |
'None' if no such job exists, a (uuid) job-id if the job is currently running or has run in the past |
Source code in kiara/registries/jobs/__init__.py
def find_matching_job_record(
self, inputs_manifest: InputsManifest
) -> Optional[uuid.UUID]:
"""Check if a job with same inputs manifest already ran some time before.
Arguments:
inputs_manifest: the manifest incl. inputs
Returns:
'None' if no such job exists, a (uuid) job-id if the job is currently running or has run in the past
"""
if inputs_manifest.job_hash in self._active_jobs.keys():
logger.debug("job.use_running")
return self._active_jobs[inputs_manifest.job_hash]
if inputs_manifest.job_hash in self._finished_jobs.keys():
job_id = self._finished_jobs[inputs_manifest.job_hash]
return job_id
matches = []
for store_id, archive in self._job_archives.items():
match = archive.find_matching_job_record(inputs_manifest=inputs_manifest)
if match:
matches.append(match)
if len(matches) == 0:
return None
elif len(matches) > 1:
raise Exception(
f"Multiple stores have a record for inputs manifest '{inputs_manifest}', this is not supported (yet)."
)
job_record = matches[0]
self._finished_jobs[inputs_manifest.job_hash] = job_record.job_id
self._archived_records[job_record.job_id] = job_record
logger.debug("job.use_cached")
return job_record.job_id
get_active_job(self, job_id)
¶Source code in kiara/registries/jobs/__init__.py
def get_active_job(self, job_id: uuid.UUID) -> ActiveJob:
if job_id in self._active_jobs.keys() or job_id in self._failed_jobs.keys():
return self._processor.get_job(job_id)
else:
if job_id in self._archived_records.keys():
raise Exception(
f"Can't retrieve active job with id '{job_id}': job is archived."
)
else:
raise Exception(f"Can't retrieve job with id '{job_id}': no such job.")
get_archive(self, store_id=None)
¶Source code in kiara/registries/jobs/__init__.py
def get_archive(self, store_id: Optional[str] = None) -> JobArchive:
if store_id is None:
store_id = self.default_job_store
if store_id is None:
raise Exception("Can't retrieve deafult job archive, none set (yet).")
return self._job_archives[store_id]
get_job_record_in_session(self, job_id)
¶Source code in kiara/registries/jobs/__init__.py
def get_job_record_in_session(self, job_id: uuid.UUID) -> JobRecord:
return self._processor.get_job_record(job_id)
get_job_status(self, job_id)
¶Source code in kiara/registries/jobs/__init__.py
def get_job_status(self, job_id: uuid.UUID) -> JobStatus:
if job_id in self._archived_records.keys():
return JobStatus.SUCCESS
elif job_id in self._failed_jobs.values():
return JobStatus.FAILED
return self._processor.get_job_status(job_id=job_id)
job_status_changed(self, job_id, old_status, new_status)
¶Source code in kiara/registries/jobs/__init__.py
def job_status_changed(
self, job_id: uuid.UUID, old_status: Optional[JobStatus], new_status: JobStatus
):
# print(f"JOB STATUS CHANGED: {job_id} - {old_status} - {new_status.value}")
if job_id in self._active_jobs.values() and new_status is JobStatus.FAILED:
job_hash = self._active_jobs.inverse.pop(job_id)
self._failed_jobs[job_hash] = job_id
elif job_id in self._active_jobs.values() and new_status is JobStatus.SUCCESS:
job_hash = self._active_jobs.inverse.pop(job_id)
job_record = self._processor.get_job_record(job_id)
self._finished_jobs[job_hash] = job_id
self._archived_records[job_id] = job_record
prepare_job_config(self, manifest, inputs)
¶Source code in kiara/registries/jobs/__init__.py
def prepare_job_config(
self, manifest: Manifest, inputs: Mapping[str, Any]
) -> JobConfig:
module = self._kiara.create_module(manifest=manifest)
job_config = JobConfig.create_from_module(
data_registry=self._kiara.data_registry, module=module, inputs=inputs
)
return job_config
register_job_archive(self, archive, alias=None)
¶Source code in kiara/registries/jobs/__init__.py
def register_job_archive(self, archive: JobArchive, alias: Optional[str] = None):
if alias is None:
alias = str(archive.archive_id)
if alias in self._job_archives.keys():
raise Exception(
f"Can't register job store, store id already registered: {alias}."
)
self._job_archives[alias] = archive
is_store = False
is_default_store = False
if isinstance(archive, JobStore):
is_store = True
if self._default_job_store is None:
self._default_job_store = alias
event = JobArchiveAddedEvent.construct(
kiara_id=self._kiara.id,
job_archive_id=archive.archive_id,
job_archive_alias=alias,
is_store=is_store,
is_default_store=is_default_store,
)
self._event_callback(event)
retrieve_result(self, job_id)
¶Source code in kiara/registries/jobs/__init__.py
def retrieve_result(self, job_id: uuid.UUID) -> ValueMap:
if job_id not in self._archived_records.keys():
self._processor.wait_for(job_id)
job_record = self._archived_records[job_id]
results = self._kiara.data_registry.load_values(job_record.outputs)
return results
store_job_record(self, job_id)
¶Source code in kiara/registries/jobs/__init__.py
def store_job_record(self, job_id: uuid.UUID):
if job_id not in self._archived_records.keys():
raise Exception(
f"Can't store job with id '{job_id}': no job record with that id exists."
)
job_record = self._archived_records[job_id]
if job_record._is_stored:
logger.debug(
"ignore.store.job_record", reason="already stored", job_id=str(job_id)
)
return
logger.debug(
"store.job_record",
job_hash=job_record.job_hash,
module_type=job_record.module_type,
)
store: JobStore = self.get_archive() # type: ignore
if not isinstance(store, JobStore):
raise Exception("Can't store job record to archive: not writable.")
pre_store_event = JobRecordPreStoreEvent.construct(
kiara_id=self._kiara.id, job_record=job_record
)
self._event_callback(pre_store_event)
store.store_job_record(job_record)
stored_event = JobRecordStoredEvent.construct(
kiara_id=self._kiara.id, job_record=job_record
)
self._event_callback(stored_event)
suppoerted_event_types(self)
¶Source code in kiara/registries/jobs/__init__.py
def suppoerted_event_types(self) -> Iterable[Type[KiaraEvent]]:
return [JobArchiveAddedEvent, JobRecordPreStoreEvent, JobRecordStoredEvent]
wait_for(self, *job_id)
¶Source code in kiara/registries/jobs/__init__.py
def wait_for(self, *job_id: uuid.UUID):
not_finished = (j for j in job_id if j not in self._archived_records.keys())
if not_finished:
self._processor.wait_for(*not_finished)
JobStore (JobArchive)
¶Source code in kiara/registries/jobs/__init__.py
class JobStore(JobArchive):
@abc.abstractmethod
def store_job_record(self, job_record: JobRecord):
pass
store_job_record(self, job_record)
¶Source code in kiara/registries/jobs/__init__.py
@abc.abstractmethod
def store_job_record(self, job_record: JobRecord):
pass
Modules¶
job_store
special
¶filesystem_store
¶
FileSystemJobArchive (JobArchive)
¶Source code in kiara/registries/jobs/job_store/filesystem_store.py
class FileSystemJobArchive(JobArchive):
_archive_type_name = "filesystem_job_archive"
_config_cls = FileSystemArchiveConfig
@classmethod
def is_writeable(cls) -> bool:
return False
@classmethod
def supported_item_types(cls) -> Iterable[str]:
return ["job_record"]
def __init__(self, archive_id: uuid.UUID, config: FileSystemArchiveConfig):
super().__init__(archive_id=archive_id, config=config)
self._base_path: Optional[Path] = None
@property
def job_store_path(self) -> Path:
if self._base_path is not None:
return self._base_path
self._base_path = Path(self.config.base_path) / str(self.archive_id)
self._base_path.mkdir(parents=True, exist_ok=True)
return self._base_path
def find_matching_job_record(
self, inputs_manifest: InputsManifest
) -> Optional[JobRecord]:
manifest_hash = inputs_manifest.manifest_hash
jobs_hash = inputs_manifest.job_hash
base_path = self.job_store_path / MANIFEST_SUB_PATH
manifest_folder = base_path / str(manifest_hash)
if not manifest_folder.exists():
return None
manifest_file = manifest_folder / "manifest.json"
if not manifest_file.exists():
raise Exception(
f"No 'manifests.json' file for manifest with hash: {manifest_hash}"
)
details_folder = manifest_folder / str(jobs_hash)
if not details_folder.exists():
return None
details_file_name = details_folder / "details.json"
if not details_file_name.exists():
raise Exception(
f"No 'inputs.json' file for manifest/inputs hash-combo: {manifest_hash} / {jobs_hash}"
)
details_content = details_file_name.read_text()
details: Dict[str, Any] = orjson.loads(details_content)
job_record = JobRecord(**details)
job_record._is_stored = True
return job_record
job_store_path: Path
property
readonly
¶
_config_cls (ArchiveConfig)
private
pydantic-model
¶find_matching_job_record(self, inputs_manifest)
¶Source code in kiara/registries/jobs/job_store/filesystem_store.py
def find_matching_job_record(
self, inputs_manifest: InputsManifest
) -> Optional[JobRecord]:
manifest_hash = inputs_manifest.manifest_hash
jobs_hash = inputs_manifest.job_hash
base_path = self.job_store_path / MANIFEST_SUB_PATH
manifest_folder = base_path / str(manifest_hash)
if not manifest_folder.exists():
return None
manifest_file = manifest_folder / "manifest.json"
if not manifest_file.exists():
raise Exception(
f"No 'manifests.json' file for manifest with hash: {manifest_hash}"
)
details_folder = manifest_folder / str(jobs_hash)
if not details_folder.exists():
return None
details_file_name = details_folder / "details.json"
if not details_file_name.exists():
raise Exception(
f"No 'inputs.json' file for manifest/inputs hash-combo: {manifest_hash} / {jobs_hash}"
)
details_content = details_file_name.read_text()
details: Dict[str, Any] = orjson.loads(details_content)
job_record = JobRecord(**details)
job_record._is_stored = True
return job_record
is_writeable()
classmethod
¶Source code in kiara/registries/jobs/job_store/filesystem_store.py
@classmethod
def is_writeable(cls) -> bool:
return False
supported_item_types()
classmethod
¶Source code in kiara/registries/jobs/job_store/filesystem_store.py
@classmethod
def supported_item_types(cls) -> Iterable[str]:
return ["job_record"]
FileSystemJobStore (FileSystemJobArchive, JobStore)
¶Source code in kiara/registries/jobs/job_store/filesystem_store.py
class FileSystemJobStore(FileSystemJobArchive, JobStore):
_archive_type_name = "filesystem_job_store"
@classmethod
def is_writeable(cls) -> bool:
return False
def store_job_record(self, job_record: JobRecord):
manifest_hash = job_record.manifest_hash
jobs_hash = job_record.job_hash
base_path = self.job_store_path / MANIFEST_SUB_PATH
manifest_folder = base_path / str(manifest_hash)
manifest_folder.mkdir(parents=True, exist_ok=True)
manifest_info_file = manifest_folder / "manifest.json"
if not manifest_info_file.exists():
manifest_info_file.write_text(job_record.manifest_data_as_json())
job_folder = manifest_folder / str(jobs_hash)
job_folder.mkdir(parents=True, exist_ok=True)
job_details_file_name = job_folder / "details.json"
if job_details_file_name.exists():
raise Exception(
f"Job record already exists: {job_details_file_name.as_posix()}"
)
job_details_file_name.write_text(job_record.json())
for output_name, output_v_id in job_record.outputs.items():
outputs_file_name = (
job_folder / f"output__{output_name}__value_id__{output_v_id}.json"
)
if outputs_file_name.exists():
# if value.pedigree_output_name == "__void__":
# return
# else:
raise Exception(f"Can't write value '{output_v_id}': already exists.")
else:
outputs_file_name.touch()
is_writeable()
classmethod
¶Source code in kiara/registries/jobs/job_store/filesystem_store.py
@classmethod
def is_writeable(cls) -> bool:
return False
store_job_record(self, job_record)
¶Source code in kiara/registries/jobs/job_store/filesystem_store.py
def store_job_record(self, job_record: JobRecord):
manifest_hash = job_record.manifest_hash
jobs_hash = job_record.job_hash
base_path = self.job_store_path / MANIFEST_SUB_PATH
manifest_folder = base_path / str(manifest_hash)
manifest_folder.mkdir(parents=True, exist_ok=True)
manifest_info_file = manifest_folder / "manifest.json"
if not manifest_info_file.exists():
manifest_info_file.write_text(job_record.manifest_data_as_json())
job_folder = manifest_folder / str(jobs_hash)
job_folder.mkdir(parents=True, exist_ok=True)
job_details_file_name = job_folder / "details.json"
if job_details_file_name.exists():
raise Exception(
f"Job record already exists: {job_details_file_name.as_posix()}"
)
job_details_file_name.write_text(job_record.json())
for output_name, output_v_id in job_record.outputs.items():
outputs_file_name = (
job_folder / f"output__{output_name}__value_id__{output_v_id}.json"
)
if outputs_file_name.exists():
# if value.pedigree_output_name == "__void__":
# return
# else:
raise Exception(f"Can't write value '{output_v_id}': already exists.")
else:
outputs_file_name.touch()
modules
special
¶
Base module for code that handles the import and management of [KiaraModule][kiara.module.KiaraModule] sub-classes.
logget
¶Classes¶
ModuleRegistry
¶Source code in kiara/registries/modules/__init__.py
class ModuleRegistry(object):
def __init__(self):
self._cached_modules: Dict[str, Dict[int, KiaraModule]] = {}
from kiara.utils.class_loading import find_all_kiara_modules
module_classes = find_all_kiara_modules()
self._module_classes: Mapping[str, Type[KiaraModule]] = {}
self._module_class_metadata: Dict[str, KiaraModuleTypeInfo] = {}
for k, v in module_classes.items():
self._module_classes[k] = v
@property
def module_types(self) -> Mapping[str, Type["KiaraModule"]]:
return self._module_classes
def get_module_class(self, module_type: str) -> Type["KiaraModule"]:
cls = self._module_classes.get(module_type, None)
if cls is None:
raise ValueError(f"No module of type '{module_type}' available.")
return cls
def get_module_type_names(self) -> Iterable[str]:
return self._module_classes.keys()
def get_module_type_metadata(self, type_name: str) -> KiaraModuleTypeInfo:
md = self._module_class_metadata.get(type_name, None)
if md is None:
md = KiaraModuleTypeInfo.create_from_type_class(
self.get_module_class(module_type=type_name)
)
self._module_class_metadata[type_name] = md
return self._module_class_metadata[type_name]
def get_context_metadata(
self, alias: Optional[str] = None, only_for_package: Optional[str] = None
) -> ModuleTypeClassesInfo:
result = {}
for type_name in self.module_types.keys():
md = self.get_module_type_metadata(type_name=type_name)
if only_for_package:
if md.context.labels.get("package") == only_for_package:
result[type_name] = md
else:
result[type_name] = md
return ModuleTypeClassesInfo.construct(group_alias=alias, type_infos=result) # type: ignore
def create_module(self, manifest: Union[Manifest, str]) -> "KiaraModule":
"""Create a [KiaraModule][kiara.module.KiaraModule] object from a module configuration.
Arguments:
manifest: the module configuration
"""
if isinstance(manifest, str):
manifest = Manifest.construct(module_type=manifest, module_config={})
if self._cached_modules.setdefault(manifest.module_type, {}).get(
manifest.manifest_hash, None
):
return self._cached_modules[manifest.module_type][manifest.manifest_hash]
if manifest.module_type in self.get_module_type_names():
m_cls: Type[KiaraModule] = self.get_module_class(manifest.module_type)
m_hash = m_cls._calculate_module_hash(manifest.module_config)
kiara_module = m_cls(module_config=manifest.module_config)
assert (
kiara_module.module_instance_hash == m_hash
) # TODO: might not be necessary? Leaving it in here for now, to see if it triggers at any stage.
else:
raise Exception(
f"Invalid module type '{manifest.module_type}'. Available type names: {', '.join(self.get_module_type_names())}"
)
return kiara_module
module_types: Mapping[str, Type[KiaraModule]]
property
readonly
¶create_module(self, manifest)
¶Create a [KiaraModule][kiara.module.KiaraModule] object from a module configuration.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
manifest |
Union[kiara.models.module.manifest.Manifest, str] |
the module configuration |
required |
Source code in kiara/registries/modules/__init__.py
def create_module(self, manifest: Union[Manifest, str]) -> "KiaraModule":
"""Create a [KiaraModule][kiara.module.KiaraModule] object from a module configuration.
Arguments:
manifest: the module configuration
"""
if isinstance(manifest, str):
manifest = Manifest.construct(module_type=manifest, module_config={})
if self._cached_modules.setdefault(manifest.module_type, {}).get(
manifest.manifest_hash, None
):
return self._cached_modules[manifest.module_type][manifest.manifest_hash]
if manifest.module_type in self.get_module_type_names():
m_cls: Type[KiaraModule] = self.get_module_class(manifest.module_type)
m_hash = m_cls._calculate_module_hash(manifest.module_config)
kiara_module = m_cls(module_config=manifest.module_config)
assert (
kiara_module.module_instance_hash == m_hash
) # TODO: might not be necessary? Leaving it in here for now, to see if it triggers at any stage.
else:
raise Exception(
f"Invalid module type '{manifest.module_type}'. Available type names: {', '.join(self.get_module_type_names())}"
)
return kiara_module
get_context_metadata(self, alias=None, only_for_package=None)
¶Source code in kiara/registries/modules/__init__.py
def get_context_metadata(
self, alias: Optional[str] = None, only_for_package: Optional[str] = None
) -> ModuleTypeClassesInfo:
result = {}
for type_name in self.module_types.keys():
md = self.get_module_type_metadata(type_name=type_name)
if only_for_package:
if md.context.labels.get("package") == only_for_package:
result[type_name] = md
else:
result[type_name] = md
return ModuleTypeClassesInfo.construct(group_alias=alias, type_infos=result) # type: ignore
get_module_class(self, module_type)
¶Source code in kiara/registries/modules/__init__.py
def get_module_class(self, module_type: str) -> Type["KiaraModule"]:
cls = self._module_classes.get(module_type, None)
if cls is None:
raise ValueError(f"No module of type '{module_type}' available.")
return cls
get_module_type_metadata(self, type_name)
¶Source code in kiara/registries/modules/__init__.py
def get_module_type_metadata(self, type_name: str) -> KiaraModuleTypeInfo:
md = self._module_class_metadata.get(type_name, None)
if md is None:
md = KiaraModuleTypeInfo.create_from_type_class(
self.get_module_class(module_type=type_name)
)
self._module_class_metadata[type_name] = md
return self._module_class_metadata[type_name]
get_module_type_names(self)
¶Source code in kiara/registries/modules/__init__.py
def get_module_type_names(self) -> Iterable[str]:
return self._module_classes.keys()
operations
special
¶
logger
¶
OperationRegistry
¶Source code in kiara/registries/operations/__init__.py
class OperationRegistry(object):
def __init__(
self,
kiara: "Kiara",
operation_type_classes: Optional[Mapping[str, Type[OperationType]]] = None,
):
self._kiara: "Kiara" = kiara
self._operation_type_classes: Optional[Dict[str, Type["OperationType"]]] = None
if operation_type_classes is not None:
self._operation_type_classes = dict(operation_type_classes)
self._operation_type_metadata: Dict[str, OperationTypeInfo] = {}
self._operation_types: Optional[Dict[str, OperationType]] = None
self._operations: Optional[Dict[str, Operation]] = None
self._operations_by_type: Optional[Dict[str, Iterable[str]]] = None
@property
def is_initialized(self) -> bool:
return self._operations is not None
@property
def operation_types(self) -> Mapping[str, OperationType]:
if self._operation_types is not None:
return self._operation_types
# TODO: support op type config
_operation_types = {}
for op_name, op_cls in self.operation_type_classes.items():
try:
_operation_types[op_name] = op_cls(
kiara=self._kiara, op_type_name=op_name
)
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
logger.debug("ignore.operation_type", operation_name=op_name, reason=e)
self._operation_types = _operation_types
return self._operation_types
def get_operation_type(self, op_type: str) -> OperationType:
if op_type not in self.operation_types.keys():
raise Exception(
f"No operation type '{op_type}' registered. Available operation types: {', '.join(self.operation_types.keys())}."
)
return self.operation_types[op_type]
def get_type_metadata(self, type_name: str) -> OperationTypeInfo:
md = self._operation_type_metadata.get(type_name, None)
if md is None:
md = OperationTypeInfo.create_from_type_class(
type_cls=self.operation_type_classes[type_name]
)
self._operation_type_metadata[type_name] = md
return self._operation_type_metadata[type_name]
def get_context_metadata(
self, alias: Optional[str] = None, only_for_package: Optional[str] = None
) -> OperationTypeClassesInfo:
result = {}
for type_name in self.operation_type_classes.keys():
md = self.get_type_metadata(type_name=type_name)
if only_for_package:
if md.context.labels.get("package") == only_for_package:
result[type_name] = md
else:
result[type_name] = md
return OperationTypeClassesInfo.construct(group_alias=alias, type_infos=result) # type: ignore
@property
def operation_type_classes(
self,
) -> Mapping[str, Type["OperationType"]]:
if self._operation_type_classes is not None:
return self._operation_type_classes
from kiara.utils.class_loading import find_all_operation_types
self._operation_type_classes = find_all_operation_types()
return self._operation_type_classes
# @property
# def operation_ids(self) -> List[str]:
# return list(self.profiles.keys())
@property
def operation_ids(self) -> Iterable[str]:
return self.operations.keys()
@property
def operations(self) -> Mapping[str, Operation]:
if self._operations is not None:
return self._operations
all_op_configs: Set[OperationConfig] = set()
for op_type in self.operation_types.values():
included_ops = op_type.retrieve_included_operation_configs()
for op in included_ops:
if isinstance(op, Mapping):
op = ManifestOperationConfig(**op)
all_op_configs.add(op)
for data_type in self._kiara.data_type_classes.values():
if hasattr(data_type, "retrieve_included_operations"):
for op in all_op_configs:
if isinstance(op, Mapping):
op = ManifestOperationConfig(**op)
all_op_configs.add(op)
operations: Dict[str, Operation] = {}
operations_by_type: Dict[str, List[str]] = {}
deferred_module_names: Dict[str, List[OperationConfig]] = {}
# first iteration
for op_config in all_op_configs:
try:
if isinstance(op_config, PipelineOperationConfig):
for mt in op_config.required_module_types:
if mt not in self._kiara.module_type_names:
deferred_module_names.setdefault(mt, []).append(op_config)
deferred_module_names.setdefault(op_config.pipeline_id, []).append(
op_config
)
continue
module_type = op_config.retrieve_module_type(kiara=self._kiara)
if module_type not in self._kiara.module_type_names:
deferred_module_names.setdefault(module_type, []).append(op_config)
else:
module_config = op_config.retrieve_module_config(kiara=self._kiara)
manifest = Manifest.construct(
module_type=module_type, module_config=module_config
)
ops = self._create_operations(manifest=manifest, doc=op_config.doc)
for op_type_name, _op in ops.items():
if _op.operation_id in operations.keys():
raise Exception(
f"Duplicate operation id: {_op.operation_id}"
)
operations[_op.operation_id] = _op
operations_by_type.setdefault(op_type_name, []).append(
_op.operation_id
)
except Exception as e:
details: Dict[str, Any] = {}
module_id = op_config.retrieve_module_type(kiara=self._kiara)
details["module_id"] = module_id
if module_id == "pipeline":
details["pipeline_id"] = op_config.pipeline_id # type: ignore
msg: Union[str, Exception] = str(e)
if not msg:
msg = e
details["details"] = msg
logger.error("invalid.operation", **details)
if is_debug():
import traceback
traceback.print_exc()
continue
error_details = {}
while deferred_module_names:
deferred_length = len(deferred_module_names)
remove_deferred_names = set()
for missing_op_id in deferred_module_names.keys():
if missing_op_id in operations.keys():
remove_deferred_names.add(missing_op_id)
continue
for op_config in deferred_module_names[missing_op_id]:
try:
if isinstance(op_config, PipelineOperationConfig):
if all(
mt in self._kiara.module_type_names
or mt in operations.keys()
for mt in op_config.required_module_types
):
module_map = {}
for mt in op_config.required_module_types:
if mt in operations.keys():
module_map[mt] = {
"module_type": operations[mt].module_type,
"module_config": operations[
mt
].module_config,
}
op_config.module_map.update(module_map)
module_config = op_config.retrieve_module_config(
kiara=self._kiara
)
manifest = Manifest.construct(
module_type="pipeline",
module_config=module_config,
)
ops = self._create_operations(
manifest=manifest, doc=op_config.doc
)
else:
missing = (
mt
for mt in op_config.required_module_types
if mt not in self._kiara.module_type_names
and mt not in operations.keys()
)
raise Exception(
f"Can't find all required module types when processing pipeline '{missing_op_id}': {', '.join(missing)}"
)
else:
raise NotImplementedError(
f"Invalid type: {type(op_config)}"
)
# module_type = op_config.retrieve_module_type(kiara=self._kiara)
# module_config = op_config.retrieve_module_config(kiara=self._kiara)
#
# # TODO: merge dicts instead of update?
# new_module_config = dict(base_config)
# new_module_config.update(module_config)
#
# manifest = Manifest.construct(module_type=operation.module_type,
# module_config=new_module_config)
for op_type_name, _op in ops.items():
if _op.operation_id in operations.keys():
raise Exception(
f"Duplicate operation id: {_op.operation_id}"
)
operations[_op.operation_id] = _op
operations_by_type.setdefault(op_type_name, []).append(
_op.operation_id
)
assert _op.operation_id == op_config.pipeline_id
for _op_id in deferred_module_names.keys():
if op_config in deferred_module_names[_op_id]:
deferred_module_names[_op_id].remove(op_config)
except Exception as e:
details = {}
module_id = op_config.retrieve_module_type(kiara=self._kiara)
details["module_id"] = module_id
if module_id == "pipeline":
details["pipeline_id"] = op_config.pipeline_id # type: ignore
msg = str(e)
if not msg:
msg = e
details["details"] = msg
error_details[missing_op_id] = details
exc_info = sys.exc_info()
details["exception"] = exc_info
continue
for name, dependencies in deferred_module_names.items():
if not dependencies:
remove_deferred_names.add(name)
for rdn in remove_deferred_names:
deferred_module_names.pop(rdn)
if len(deferred_module_names) == deferred_length:
for mn in deferred_module_names:
if mn in operations.keys():
continue
details = error_details.get(missing_op_id, {"details": "-- n/a --"})
exception = details.pop("exception", None)
if exception and is_debug():
import traceback
traceback.print_exception(*exception)
logger.error(f"invalid.operation.{mn}", operation_id=mn, **details)
break
self._operations = {}
for missing_op_id in sorted(operations.keys()):
self._operations[missing_op_id] = operations[missing_op_id]
self._operations_by_type = {}
for op_type_name in sorted(operations_by_type.keys()):
self._operations_by_type.setdefault(
op_type_name, sorted(operations_by_type[op_type_name])
)
return self._operations
def _create_operations(self, manifest: Manifest, doc: Any) -> Dict[str, Operation]:
module = self._kiara.create_module(manifest)
op_types = {}
for op_name, op_type in self.operation_types.items():
op_details = op_type.check_matching_operation(module=module)
if not op_details:
continue
operation = Operation(
module_type=manifest.module_type,
module_config=manifest.module_config,
operation_id=op_details.operation_id,
operation_details=op_details,
module_details=KiaraModuleClass.from_module(module),
doc=doc,
)
operation._module = module
op_types[op_name] = operation
return op_types
def get_operation(self, operation_id: str) -> Operation:
if operation_id not in self.operation_ids:
raise Exception(f"No operation registered with id: {operation_id}")
op = self.operations[operation_id]
return op
def find_all_operation_types(self, operation_id: str) -> Set[str]:
result = set()
for op_type, ops in self.operations_by_type.items():
if operation_id in ops:
result.add(op_type)
return result
@property
def operations_by_type(self) -> Mapping[str, Iterable[str]]:
if self._operations_by_type is None:
self.operations # noqa
return self._operations_by_type # type: ignore
is_initialized: bool
property
readonly
¶operation_ids: Iterable[str]
property
readonly
¶operation_type_classes: Mapping[str, Type[OperationType]]
property
readonly
¶operation_types: Mapping[str, kiara.operations.OperationType]
property
readonly
¶operations: Mapping[str, kiara.models.module.operation.Operation]
property
readonly
¶operations_by_type: Mapping[str, Iterable[str]]
property
readonly
¶find_all_operation_types(self, operation_id)
¶Source code in kiara/registries/operations/__init__.py
def find_all_operation_types(self, operation_id: str) -> Set[str]:
result = set()
for op_type, ops in self.operations_by_type.items():
if operation_id in ops:
result.add(op_type)
return result
get_context_metadata(self, alias=None, only_for_package=None)
¶Source code in kiara/registries/operations/__init__.py
def get_context_metadata(
self, alias: Optional[str] = None, only_for_package: Optional[str] = None
) -> OperationTypeClassesInfo:
result = {}
for type_name in self.operation_type_classes.keys():
md = self.get_type_metadata(type_name=type_name)
if only_for_package:
if md.context.labels.get("package") == only_for_package:
result[type_name] = md
else:
result[type_name] = md
return OperationTypeClassesInfo.construct(group_alias=alias, type_infos=result) # type: ignore
get_operation(self, operation_id)
¶Source code in kiara/registries/operations/__init__.py
def get_operation(self, operation_id: str) -> Operation:
if operation_id not in self.operation_ids:
raise Exception(f"No operation registered with id: {operation_id}")
op = self.operations[operation_id]
return op
get_operation_type(self, op_type)
¶Source code in kiara/registries/operations/__init__.py
def get_operation_type(self, op_type: str) -> OperationType:
if op_type not in self.operation_types.keys():
raise Exception(
f"No operation type '{op_type}' registered. Available operation types: {', '.join(self.operation_types.keys())}."
)
return self.operation_types[op_type]
get_type_metadata(self, type_name)
¶Source code in kiara/registries/operations/__init__.py
def get_type_metadata(self, type_name: str) -> OperationTypeInfo:
md = self._operation_type_metadata.get(type_name, None)
if md is None:
md = OperationTypeInfo.create_from_type_class(
type_cls=self.operation_type_classes[type_name]
)
self._operation_type_metadata[type_name] = md
return self._operation_type_metadata[type_name]
types
special
¶
TYPE_PROFILE_MAP
¶Classes¶
TypeRegistry
¶Source code in kiara/registries/types/__init__.py
class TypeRegistry(object):
def __init__(self, kiara: "Kiara"):
self._kiara: Kiara = kiara
self._data_types: Optional[bidict[str, Type[DataType]]] = None
self._data_type_metadata: Dict[str, DataTypeClassInfo] = {}
self._cached_data_type_objects: Dict[int, DataType] = {}
# self._registered_python_classes: Dict[Type, typing.List[str]] = None # type: ignore
self._type_hierarchy: Optional[nx.DiGraph] = None
self._lineages_cache: Dict[str, List[str]] = {}
self._type_profiles: Optional[Dict[str, Mapping[str, Any]]] = None
def invalidate_types(self):
self._data_types = None
# self._registered_python_classes = None
def retrieve_data_type(
self, data_type_name: str, data_type_config: Optional[Mapping[str, Any]] = None
) -> DataType:
if data_type_config is None:
data_type_config = {}
else:
data_type_config = dict(data_type_config)
if data_type_name not in self.data_type_profiles.keys():
raise Exception(f"Data type name not registered: {data_type_name}")
data_type: str = self.data_type_profiles[data_type_name]["type_name"]
type_config = self.data_type_profiles[data_type_name]["type_config"]
if data_type_config:
type_config = dict(type_config)
type_config.update(data_type_config)
cls = self.get_data_type_cls(type_name=data_type)
hash = cls._calculate_data_type_hash(type_config)
if hash in self._cached_data_type_objects.keys():
return self._cached_data_type_objects[hash]
result = cls(**type_config)
assert result.data_type_hash == hash
self._cached_data_type_objects[result.data_type_hash] = result
return result
@property
def data_type_classes(self) -> bidict[str, Type[DataType]]:
if self._data_types is not None:
return self._data_types
self._data_types = bidict(find_all_data_types())
profiles: Dict[str, Mapping[str, Any]] = {
dn: {"type_name": dn, "type_config": {}} for dn in self._data_types.keys()
}
for name, cls in self._data_types.items():
cls_profiles = cls.retrieve_available_type_profiles()
for profile_name, type_config in cls_profiles.items():
if profile_name in profiles.keys():
raise Exception(f"Duplicate data type profile: {profile_name}")
profiles[profile_name] = {"type_name": name, "type_config": type_config}
self._type_profiles = profiles
return self._data_types
@property
def data_type_profiles(self) -> Mapping[str, Mapping[str, Any]]:
if self._type_profiles is None:
self.data_type_classes # noqa
assert self._type_profiles is not None
return self._type_profiles
@property
def data_type_hierarchy(self) -> "nx.DiGraph":
if self._type_hierarchy is not None:
return self._type_hierarchy
def recursive_base_find(cls: Type, current: Optional[List[str]] = None):
if current is None:
current = []
for base in cls.__bases__:
if base in self.data_type_classes.values():
current.append(self.data_type_classes.inverse[base])
recursive_base_find(base, current=current)
return current
bases = {}
for name, cls in self.data_type_classes.items():
bases[name] = recursive_base_find(cls)
for profile_name, details in self.data_type_profiles.items():
if not details["type_config"]:
continue
if profile_name in bases.keys():
raise Exception(
f"Invalid profile name '{profile_name}': shadowing data type. This is most likely a bug."
)
bases[profile_name] = [details["type_name"]]
import networkx as nx
hierarchy = nx.DiGraph()
hierarchy.add_node(KIARA_ROOT_TYPE_NAME)
for name, _bases in bases.items():
profile_details = self.data_type_profiles[name]
cls = self.data_type_classes[profile_details["type_name"]]
hierarchy.add_node(name, cls=cls)
if not _bases:
hierarchy.add_edge(KIARA_ROOT_TYPE_NAME, name)
else:
# we only need the first parent, all others will be taken care of by the parent of the parent
hierarchy.add_edge(_bases[0], name)
self._type_hierarchy = hierarchy
return self._type_hierarchy
def get_sub_hierarchy(self, data_type: str):
import networkx as nx
graph: nx.DiGraph = self.data_type_hierarchy
desc = nx.descendants(graph, data_type)
desc.add(data_type)
sub_graph = graph.subgraph(desc)
return sub_graph
def get_type_lineage(self, data_type_name: str) -> Iterable[str]:
"""Returns the shortest path between the specified type and the root, in reverse direction starting from the specified type."""
if data_type_name not in self.data_type_profiles.keys():
raise Exception(f"No data type '{data_type_name}' registered.")
if data_type_name in self._lineages_cache.keys():
return self._lineages_cache[data_type_name]
import networkx as nx
path = nx.shortest_path(
self.data_type_hierarchy, KIARA_ROOT_TYPE_NAME, data_type_name
)
path.remove(KIARA_ROOT_TYPE_NAME)
self._lineages_cache[data_type_name] = list(reversed(path))
return self._lineages_cache[data_type_name]
def get_sub_types(self, data_type_name: str) -> Set[str]:
if data_type_name not in self.data_type_classes.keys():
raise Exception(f"No data type '{data_type_name}' registered.")
import networkx as nx
desc = nx.descendants(self.data_type_hierarchy, data_type_name)
return desc
def get_associated_profiles(
self, data_type_name: str
) -> Mapping[str, Mapping[str, Any]]:
if data_type_name not in self.data_type_classes.keys():
raise Exception(f"No data type '{data_type_name}' registered.")
result = {}
for profile_name, details in self.data_type_profiles.items():
if (
profile_name != data_type_name
and data_type_name == details["type_name"]
):
result[profile_name] = details
return result
@property
def data_type_names(self) -> List[str]:
return list(self.data_type_profiles.keys())
def get_data_type_cls(self, type_name: str) -> Type[DataType]:
_type_details = self.data_type_profiles.get(type_name, None)
if _type_details is None:
raise Exception(
f"No value type '{type_name}', available types: {', '.join(self.data_type_profiles.keys())}"
)
resolved_type_name: str = _type_details["type_name"]
t = self.data_type_classes.get(resolved_type_name, None)
if t is None:
raise Exception(
f"No value type '{type_name}', available types: {', '.join(self.data_type_profiles.keys())}"
)
return t
def get_type_metadata(self, type_name: str) -> DataTypeClassInfo:
md = self._data_type_metadata.get(type_name, None)
if md is None:
md = DataTypeClassInfo.create_from_type_class(
type_cls=self.get_data_type_cls(type_name=type_name), kiara=self._kiara
)
self._data_type_metadata[type_name] = md
return self._data_type_metadata[type_name]
def get_context_metadata(
self, alias: Optional[str] = None, only_for_package: Optional[str] = None
) -> DataTypeClassesInfo:
result = {}
for type_name in self.data_type_classes.keys():
md = self.get_type_metadata(type_name=type_name)
if only_for_package:
if md.context.labels.get("package") == only_for_package:
result[type_name] = md
else:
result[type_name] = md
_result = DataTypeClassesInfo.construct(group_alias=alias, type_infos=result) # type: ignore
_result._kiara = self._kiara
return _result
data_type_classes: bidict
property
readonly
¶data_type_hierarchy: nx.DiGraph
property
readonly
¶data_type_names: List[str]
property
readonly
¶data_type_profiles: Mapping[str, Mapping[str, Any]]
property
readonly
¶get_associated_profiles(self, data_type_name)
¶Source code in kiara/registries/types/__init__.py
def get_associated_profiles(
self, data_type_name: str
) -> Mapping[str, Mapping[str, Any]]:
if data_type_name not in self.data_type_classes.keys():
raise Exception(f"No data type '{data_type_name}' registered.")
result = {}
for profile_name, details in self.data_type_profiles.items():
if (
profile_name != data_type_name
and data_type_name == details["type_name"]
):
result[profile_name] = details
return result
get_context_metadata(self, alias=None, only_for_package=None)
¶Source code in kiara/registries/types/__init__.py
def get_context_metadata(
self, alias: Optional[str] = None, only_for_package: Optional[str] = None
) -> DataTypeClassesInfo:
result = {}
for type_name in self.data_type_classes.keys():
md = self.get_type_metadata(type_name=type_name)
if only_for_package:
if md.context.labels.get("package") == only_for_package:
result[type_name] = md
else:
result[type_name] = md
_result = DataTypeClassesInfo.construct(group_alias=alias, type_infos=result) # type: ignore
_result._kiara = self._kiara
return _result
get_data_type_cls(self, type_name)
¶Source code in kiara/registries/types/__init__.py
def get_data_type_cls(self, type_name: str) -> Type[DataType]:
_type_details = self.data_type_profiles.get(type_name, None)
if _type_details is None:
raise Exception(
f"No value type '{type_name}', available types: {', '.join(self.data_type_profiles.keys())}"
)
resolved_type_name: str = _type_details["type_name"]
t = self.data_type_classes.get(resolved_type_name, None)
if t is None:
raise Exception(
f"No value type '{type_name}', available types: {', '.join(self.data_type_profiles.keys())}"
)
return t
get_sub_hierarchy(self, data_type)
¶Source code in kiara/registries/types/__init__.py
def get_sub_hierarchy(self, data_type: str):
import networkx as nx
graph: nx.DiGraph = self.data_type_hierarchy
desc = nx.descendants(graph, data_type)
desc.add(data_type)
sub_graph = graph.subgraph(desc)
return sub_graph
get_sub_types(self, data_type_name)
¶Source code in kiara/registries/types/__init__.py
def get_sub_types(self, data_type_name: str) -> Set[str]:
if data_type_name not in self.data_type_classes.keys():
raise Exception(f"No data type '{data_type_name}' registered.")
import networkx as nx
desc = nx.descendants(self.data_type_hierarchy, data_type_name)
return desc
get_type_lineage(self, data_type_name)
¶Returns the shortest path between the specified type and the root, in reverse direction starting from the specified type.
Source code in kiara/registries/types/__init__.py
def get_type_lineage(self, data_type_name: str) -> Iterable[str]:
"""Returns the shortest path between the specified type and the root, in reverse direction starting from the specified type."""
if data_type_name not in self.data_type_profiles.keys():
raise Exception(f"No data type '{data_type_name}' registered.")
if data_type_name in self._lineages_cache.keys():
return self._lineages_cache[data_type_name]
import networkx as nx
path = nx.shortest_path(
self.data_type_hierarchy, KIARA_ROOT_TYPE_NAME, data_type_name
)
path.remove(KIARA_ROOT_TYPE_NAME)
self._lineages_cache[data_type_name] = list(reversed(path))
return self._lineages_cache[data_type_name]
get_type_metadata(self, type_name)
¶Source code in kiara/registries/types/__init__.py
def get_type_metadata(self, type_name: str) -> DataTypeClassInfo:
md = self._data_type_metadata.get(type_name, None)
if md is None:
md = DataTypeClassInfo.create_from_type_class(
type_cls=self.get_data_type_cls(type_name=type_name), kiara=self._kiara
)
self._data_type_metadata[type_name] = md
return self._data_type_metadata[type_name]
invalidate_types(self)
¶Source code in kiara/registries/types/__init__.py
def invalidate_types(self):
self._data_types = None
# self._registered_python_classes = None
retrieve_data_type(self, data_type_name, data_type_config=None)
¶Source code in kiara/registries/types/__init__.py
def retrieve_data_type(
self, data_type_name: str, data_type_config: Optional[Mapping[str, Any]] = None
) -> DataType:
if data_type_config is None:
data_type_config = {}
else:
data_type_config = dict(data_type_config)
if data_type_name not in self.data_type_profiles.keys():
raise Exception(f"Data type name not registered: {data_type_name}")
data_type: str = self.data_type_profiles[data_type_name]["type_name"]
type_config = self.data_type_profiles[data_type_name]["type_config"]
if data_type_config:
type_config = dict(type_config)
type_config.update(data_type_config)
cls = self.get_data_type_cls(type_name=data_type)
hash = cls._calculate_data_type_hash(type_config)
if hash in self._cached_data_type_objects.keys():
return self._cached_data_type_objects[hash]
result = cls(**type_config)
assert result.data_type_hash == hash
self._cached_data_type_objects[result.data_type_hash] = result
return result
service
special
¶
rest
special
¶
KiaraAPI
¶Source code in kiara/service/rest/__init__.py
class KiaraAPI(object):
def __init__(self, kiara: Kiara):
self._kiara: Kiara = kiara
self._config: Optional[Config] = None
@property
def app(self):
app = FastAPI()
static_dir = os.path.join(os.path.dirname(__file__), "static")
app.mount(
"/static", StaticFiles(directory=static_dir, html=True), name="static"
)
@app.get("/context/", response_model=KiaraContextInfo)
async def context():
return self._kiara.context_info
@app.get("/operations/", response_model=OperationGroupInfo)
async def operations(test: Optional[bool] = False):
print(test)
return self._kiara.context_info.operations
@app.get("/operation/{operation_id}", response_class=HTMLResponse)
async def get_operation(operation_id):
print(operation_id)
return self._kiara.context_info.operations[operation_id].create_html()
self._app = app
return self._app
@property
def config(self) -> Config:
if self._config is not None:
return self._config
self._config = Config()
self._config.bind = ["localhost:8888"]
return self._config
def start(self):
asyncio.run(serve(self.app, self.config))
utils
special
¶
CAMEL_TO_SNAKE_REGEX
¶
SUBCLASS_TYPE
¶
WORD_REGEX_PATTERN
¶
logger
¶
string_types
¶
yaml
¶
StringYAML (YAML)
¶
Source code in kiara/utils/__init__.py
class StringYAML(YAML):
def dump(self, data, stream=None, **kw):
inefficient = False
if stream is None:
inefficient = True
stream = StringIO()
YAML.dump(self, data, stream, **kw)
if inefficient:
return stream.getvalue()
dump(self, data, stream=None, **kw)
¶Source code in kiara/utils/__init__.py
def dump(self, data, stream=None, **kw):
inefficient = False
if stream is None:
inefficient = True
stream = StringIO()
YAML.dump(self, data, stream, **kw)
if inefficient:
return stream.getvalue()
Functions¶
camel_case_to_snake_case(camel_text, repl='_')
¶
Source code in kiara/utils/__init__.py
def camel_case_to_snake_case(camel_text: str, repl: str = "_"):
return CAMEL_TO_SNAKE_REGEX.sub(repl, camel_text).lower()
check_valid_field_names(*field_names)
¶
Check whether the provided field names are all valid.
Returns:
| Type | Description |
|---|---|
List[str] |
an iterable of strings with invalid field names |
Source code in kiara/utils/__init__.py
def check_valid_field_names(*field_names) -> List[str]:
"""Check whether the provided field names are all valid.
Returns:
an iterable of strings with invalid field names
"""
return [x for x in field_names if x in INVALID_VALUE_NAMES or x.startswith("_")]
create_valid_identifier(text)
¶
Source code in kiara/utils/__init__.py
def create_valid_identifier(text: str):
return slugify(text, separator="_")
dict_from_cli_args(*args, *, list_keys=None)
¶
Source code in kiara/utils/__init__.py
def dict_from_cli_args(
*args: str, list_keys: Optional[Iterable[str]] = None
) -> Dict[str, Any]:
if not args:
return {}
config: Dict[str, Any] = {}
for arg in args:
if "=" in arg:
key, value = arg.split("=", maxsplit=1)
try:
_v = json.loads(value)
except Exception:
_v = value
part_config = {key: _v}
elif os.path.isfile(os.path.realpath(os.path.expanduser(arg))):
path = os.path.realpath(os.path.expanduser(arg))
part_config = get_data_from_file(path)
assert isinstance(part_config, Mapping)
else:
try:
part_config = json.loads(arg)
assert isinstance(part_config, Mapping)
except Exception:
raise Exception(f"Could not parse argument into data: {arg}")
if list_keys is None:
list_keys = []
for k, v in part_config.items():
if k in list_keys:
config.setdefault(k, []).append(v)
else:
if k in config.keys():
logger.warning("duplicate.key", old_value=k, new_value=v)
config[k] = v
return config
find_free_id(stem, current_ids, sep='_')
¶
Find a free var (or other name) based on a stem string, based on a list of provided existing names.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
stem |
str |
the base string to use |
required |
current_ids |
Iterable[str] |
currently existing names |
required |
method |
str |
the method to create new names (allowed: 'count' -- for now) |
required |
method_args |
dict |
prototing_config for the creation method |
required |
Returns:
| Type | Description |
|---|---|
str |
a free name |
Source code in kiara/utils/__init__.py
def find_free_id(
stem: str,
current_ids: Iterable[str],
sep="_",
) -> str:
"""Find a free var (or other name) based on a stem string, based on a list of provided existing names.
Args:
stem (str): the base string to use
current_ids (Iterable[str]): currently existing names
method (str): the method to create new names (allowed: 'count' -- for now)
method_args (dict): prototing_config for the creation method
Returns:
str: a free name
"""
start_count = 1
if stem not in current_ids:
return stem
i = start_count
# new_name = None
while True:
new_name = f"{stem}{sep}{i}"
if new_name in current_ids:
i = i + 1
continue
break
return new_name
first_line(text)
¶
Source code in kiara/utils/__init__.py
def first_line(text: str):
if "\n" in text:
return text.split("\n")[0].strip()
else:
return text
get_auto_workflow_alias(module_type, use_incremental_ids=False)
¶
Return an id for a workflow obj of a provided module class.
If 'use_incremental_ids' is set to True, a unique id is returned.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
module_type |
str |
the name of the module type |
required |
use_incremental_ids |
bool |
whether to return a unique (incremental) id |
False |
Returns:
| Type | Description |
|---|---|
str |
a module id |
Source code in kiara/utils/__init__.py
def get_auto_workflow_alias(module_type: str, use_incremental_ids: bool = False) -> str:
"""Return an id for a workflow obj of a provided module class.
If 'use_incremental_ids' is set to True, a unique id is returned.
Args:
module_type (str): the name of the module type
use_incremental_ids (bool): whether to return a unique (incremental) id
Returns:
str: a module id
"""
if not use_incremental_ids:
return module_type
nr = _AUTO_MODULE_ID.setdefault(module_type, 0)
_AUTO_MODULE_ID[module_type] = nr + 1
return f"{module_type}_{nr}"
get_data_from_file(path, content_type=None)
¶
Source code in kiara/utils/__init__.py
def get_data_from_file(
path: Union[str, Path], content_type: Optional[str] = None
) -> Any:
if isinstance(path, str):
path = Path(os.path.expanduser(path))
content = path.read_text()
if content_type:
assert content_type in ["json", "yaml"]
else:
if path.name.endswith(".json"):
content_type = "json"
elif path.name.endswith(".yaml") or path.name.endswith(".yml"):
content_type = "yaml"
else:
raise ValueError(
"Invalid data format, only 'json' or 'yaml' are supported currently."
)
if content_type == "json":
data = json.loads(content)
else:
data = yaml.load(content)
return data
is_debug()
¶
Source code in kiara/utils/__init__.py
def is_debug() -> bool:
debug = os.environ.get("DEBUG", "")
if debug.lower() == "true":
return True
else:
return False
is_develop()
¶
Source code in kiara/utils/__init__.py
def is_develop() -> bool:
debug = os.environ.get("DEVELOP", "")
if debug.lower() == "true":
return True
else:
return False
is_rich_renderable(item)
¶
Source code in kiara/utils/__init__.py
def is_rich_renderable(item: Any):
return isinstance(item, (ConsoleRenderable, RichCast, str))
log_message(msg, **data)
¶
Source code in kiara/utils/__init__.py
def log_message(msg: str, **data):
if is_debug():
logger.debug(msg, **data)
# else:
# logger.debug(msg, **data)
merge_dicts(*dicts)
¶
Source code in kiara/utils/__init__.py
def merge_dicts(*dicts: Mapping[str, Any]) -> Dict[str, Any]:
if not dicts:
return {}
current: Dict[str, Any] = {}
for d in dicts:
dpath.util.merge(current, copy.deepcopy(d))
return current
orjson_dumps(v, *, default=None, **args)
¶
Source code in kiara/utils/__init__.py
def orjson_dumps(v, *, default=None, **args):
# orjson.dumps returns bytes, to match standard json.dumps we need to decode
try:
return orjson.dumps(v, default=default, **args).decode()
except Exception as e:
if is_debug():
print(f"Error dumping json data: {e}")
from kiara import dbg
dbg(v)
raise e
to_camel_case(text)
¶
Source code in kiara/utils/__init__.py
def to_camel_case(text: str) -> str:
words = WORD_REGEX_PATTERN.split(text)
return "".join(w.title() for i, w in enumerate(words))
Modules¶
class_loading
¶
KiaraEntryPointItem
¶KiaraEntryPointIterable
¶SUBCLASS_TYPE
¶logger
¶Functions¶
find_all_archive_types()
¶Find all KiaraArchive subclasses via package entry points.
Source code in kiara/utils/class_loading.py
def find_all_archive_types() -> Dict[str, Type["KiaraArchive"]]:
"""Find all [KiaraArchive][kiara.registries.KiaraArchive] subclasses via package entry points."""
from kiara.registries import KiaraArchive
return load_all_subclasses_for_entry_point(
entry_point_name="kiara.archive_type",
base_class=KiaraArchive, # type: ignore
type_id_key="_archive_type_name",
type_id_func=_cls_name_id_func,
attach_python_metadata=False,
)
find_all_data_types()
¶Find all [KiaraModule][kiara.module.KiaraModule] subclasses via package entry points.
TODO
Source code in kiara/utils/class_loading.py
def find_all_data_types() -> Dict[str, Type["DataType"]]:
"""Find all [KiaraModule][kiara.module.KiaraModule] subclasses via package entry points.
TODO
"""
from kiara.data_types import DataType
all_data_types = load_all_subclasses_for_entry_point(
entry_point_name="kiara.data_types",
base_class=DataType, # type: ignore
type_id_key="_data_type_name",
type_id_func=_cls_name_id_func,
)
invalid = [x for x in all_data_types.keys() if "." in x]
if invalid:
raise Exception(
f"Invalid value type name(s), type names can't contain '.': {', '.join(invalid)}"
)
return all_data_types
find_all_kiara_modules()
¶Find all [KiaraModule][kiara.module.KiaraModule] subclasses via package entry points.
TODO
Source code in kiara/utils/class_loading.py
def find_all_kiara_modules() -> Dict[str, Type["KiaraModule"]]:
"""Find all [KiaraModule][kiara.module.KiaraModule] subclasses via package entry points.
TODO
"""
from kiara.modules import KiaraModule
modules = load_all_subclasses_for_entry_point(
entry_point_name="kiara.modules",
base_class=KiaraModule, # type: ignore
type_id_key="_module_type_name",
attach_python_metadata=True,
)
result = {}
# need to test this, since I couldn't add an abstract method to the KiaraModule class itself (mypy complained because it is potentially overloaded)
for k, cls in modules.items():
if not hasattr(cls, "process"):
# TODO: check signature of process method
log_message("ignore.module.class", cls=cls, reason="no 'process' method")
continue
result[k] = cls
return result
find_all_kiara_pipeline_paths(skip_errors=False)
¶Source code in kiara/utils/class_loading.py
def find_all_kiara_pipeline_paths(
skip_errors: bool = False,
) -> List[str]:
import logging
log2 = logging.getLogger("stevedore")
out_hdlr = logging.StreamHandler(sys.stdout)
out_hdlr.setFormatter(
logging.Formatter("kiara pipeline search plugin error -> %(message)s")
)
out_hdlr.setLevel(logging.INFO)
log2.addHandler(out_hdlr)
log2.setLevel(logging.INFO)
log_message("events.loading.pipelines")
mgr = ExtensionManager(
namespace="kiara.pipelines", invoke_on_load=False, propagate_map_exceptions=True
)
paths: List[str] = []
# TODO: make sure we load 'core' first?
for plugin in mgr:
name = plugin.name
if (
isinstance(plugin.plugin, tuple)
and len(plugin.plugin) >= 1
and callable(plugin.plugin[0])
) or callable(plugin.plugin):
try:
if callable(plugin.plugin):
func = plugin.plugin
args = []
else:
func = plugin.plugin[0]
args = plugin.plugin[1:]
result = func(*args)
if isinstance(result, str):
paths.append(result)
else:
paths.extend(result)
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
if skip_errors:
log_message(
"ignore.pipline_entrypoint", entrypoint_name=name, reason=str(e)
)
continue
raise Exception(f"Error trying to load plugin '{plugin.plugin}': {e}")
else:
if skip_errors:
log_message(
"ignore.pipline_entrypoint",
entrypoint_name=name,
reason=f"invalid plugin type '{type(plugin.plugin)}'",
)
continue
msg = f"Can't load pipelines for entrypoint '{name}': invalid plugin type '{type(plugin.plugin)}'"
raise Exception(msg)
return paths
find_all_operation_types()
¶Source code in kiara/utils/class_loading.py
def find_all_operation_types() -> Dict[str, Type["OperationType"]]:
from kiara.operations import OperationType
result = load_all_subclasses_for_entry_point(
entry_point_name="kiara.operation_types",
base_class=OperationType, # type: ignore
type_id_key="_operation_type_name",
)
return result
find_all_value_metadata_models()
¶Find all [KiaraModule][kiara.module.KiaraModule] subclasses via package entry points.
TODO
Source code in kiara/utils/class_loading.py
def find_all_value_metadata_models() -> Dict[str, Type["ValueMetadata"]]:
"""Find all [KiaraModule][kiara.module.KiaraModule] subclasses via package entry points.
TODO
"""
from kiara.models.values.value_metadata import ValueMetadata
return load_all_subclasses_for_entry_point(
entry_point_name="kiara.metadata_models",
base_class=ValueMetadata, # type: ignore
type_id_key="_metadata_key",
type_id_func=_cls_name_id_func,
attach_python_metadata=False,
)
find_data_types_under(module)
¶Source code in kiara/utils/class_loading.py
def find_data_types_under(module: Union[str, ModuleType]) -> List[Type["DataType"]]:
from kiara.data_types import DataType
return find_subclasses_under(
base_class=DataType, # type: ignore
python_module=module,
)
find_kiara_modules_under(module)
¶Source code in kiara/utils/class_loading.py
def find_kiara_modules_under(
module: Union[str, ModuleType],
) -> List[Type["KiaraModule"]]:
from kiara.modules import KiaraModule
return find_subclasses_under(
base_class=KiaraModule, # type: ignore
python_module=module,
)
find_operations_under(module)
¶Source code in kiara/utils/class_loading.py
def find_operations_under(
module: Union[str, ModuleType]
) -> List[Type["OperationType"]]:
from kiara.operations import OperationType
return find_subclasses_under(
base_class=OperationType, # type: ignore
python_module=module,
)
find_pipeline_base_path_for_module(module)
¶Source code in kiara/utils/class_loading.py
def find_pipeline_base_path_for_module(module: Union[str, ModuleType]) -> Optional[str]:
if hasattr(sys, "frozen"):
raise NotImplementedError("Pyinstaller bundling not supported yet.")
if isinstance(module, str):
module = importlib.import_module(module)
module_file = module.__file__
assert module_file is not None
path = os.path.dirname(module_file)
if not os.path.exists:
log_message("ignore.pipeline_folder", path=path, reason="folder does not exist")
return None
return path
find_subclasses_under(base_class, python_module)
¶Find all (non-abstract) subclasses of a base class that live under a module (recursively).
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
base_class |
Type[~SUBCLASS_TYPE] |
the parent class |
required |
python_module |
Union[str, module] |
the Python module to search |
required |
Returns:
| Type | Description |
|---|---|
List[Type[~SUBCLASS_TYPE]] |
a list of all subclasses |
Source code in kiara/utils/class_loading.py
def find_subclasses_under(
base_class: Type[SUBCLASS_TYPE],
python_module: Union[str, ModuleType],
) -> List[Type[SUBCLASS_TYPE]]:
"""Find all (non-abstract) subclasses of a base class that live under a module (recursively).
Arguments:
base_class: the parent class
python_module: the Python module to search
Returns:
a list of all subclasses
"""
if hasattr(sys, "frozen"):
raise NotImplementedError("Pyinstaller bundling not supported yet.")
if isinstance(python_module, str):
python_module = importlib.import_module(python_module)
_import_modules_recursively(python_module)
subclasses: Iterable[Type[SUBCLASS_TYPE]] = _get_all_subclasses(base_class)
result = []
for sc in subclasses:
if not sc.__module__.startswith(python_module.__name__):
continue
result.append(sc)
return result
find_value_metadata_models_under(module)
¶Source code in kiara/utils/class_loading.py
def find_value_metadata_models_under(
module: Union[str, ModuleType]
) -> List[Type["ValueMetadata"]]:
from kiara.models.values.value_metadata import ValueMetadata
result = find_subclasses_under(
base_class=ValueMetadata, # type: ignore
python_module=module,
)
return result
load_all_subclasses_for_entry_point(entry_point_name, base_class, ignore_abstract_classes=True, type_id_key=None, type_id_func=None, type_id_no_attach=False, attach_python_metadata=False)
¶Find all subclasses of a base class via package entry points.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
entry_point_name |
str |
the entry point name to query entries for |
required |
base_class |
Type[~SUBCLASS_TYPE] |
the base class to look for |
required |
ignore_abstract_classes |
bool |
whether to include abstract classes in the result |
True |
type_id_key |
Optional[str] |
if provided, the found classes will have their id attached as an attribute, using the value of this as the name. if an attribute of this name already exists, it will be used as id without further processing |
None |
type_id_func |
Callable |
a function to take the found class as input, and returns a string representing the id of the class. By default, the module path + "." + class name (snake-case) is used (minus the string 'kiara_modules. |
None |
type_id_no_attach |
bool |
in case you want to use the type_id_key to set the id, but don't want it attached to classes that don't have it, set this to true. In most cases, you won't need this option |
False |
attach_python_metadata |
Union[bool, str] |
whether to attach a PythonClass metadata model to the class. By default, '_python_class' is used as attribute name if this argument is 'True', If this argument is a string, that will be used as name instead. |
False |
Source code in kiara/utils/class_loading.py
def load_all_subclasses_for_entry_point(
entry_point_name: str,
base_class: Type[SUBCLASS_TYPE],
ignore_abstract_classes: bool = True,
type_id_key: Optional[str] = None,
type_id_func: Callable = None,
type_id_no_attach: bool = False,
attach_python_metadata: Union[bool, str] = False,
) -> Dict[str, Type[SUBCLASS_TYPE]]:
"""Find all subclasses of a base class via package entry points.
Arguments:
entry_point_name: the entry point name to query entries for
base_class: the base class to look for
ignore_abstract_classes: whether to include abstract classes in the result
type_id_key: if provided, the found classes will have their id attached as an attribute, using the value of this as the name. if an attribute of this name already exists, it will be used as id without further processing
type_id_func: a function to take the found class as input, and returns a string representing the id of the class. By default, the module path + "." + class name (snake-case) is used (minus the string 'kiara_modules.<project_name>'', if it exists at the beginning
type_id_no_attach: in case you want to use the type_id_key to set the id, but don't want it attached to classes that don't have it, set this to true. In most cases, you won't need this option
attach_python_metadata: whether to attach a [PythonClass][kiara.models.python_class.PythonClass] metadata model to the class. By default, '_python_class' is used as attribute name if this argument is 'True', If this argument is a string, that will be used as name instead.
"""
log2 = logging.getLogger("stevedore")
out_hdlr = logging.StreamHandler(sys.stdout)
out_hdlr.setFormatter(
logging.Formatter(
f"{entry_point_name} plugin search message/error -> %(message)s"
)
)
out_hdlr.setLevel(logging.INFO)
log2.addHandler(out_hdlr)
if is_debug():
log2.setLevel(logging.DEBUG)
else:
out_hdlr.setLevel(logging.INFO)
log2.setLevel(logging.INFO)
log_message("events.loading.entry_points", entry_point_name=entry_point_name)
mgr = ExtensionManager(
namespace=entry_point_name,
invoke_on_load=False,
propagate_map_exceptions=True,
)
result_entrypoints: Dict[str, Type[SUBCLASS_TYPE]] = {}
result_dynamic: Dict[str, Type[SUBCLASS_TYPE]] = {}
for plugin in mgr:
name = plugin.name
if isinstance(plugin.plugin, type):
# this means an actual (sub-)class was provided in the entrypoint
cls = plugin.plugin
if not issubclass(cls, base_class):
log_message(
"ignore.entrypoint",
entry_point=name,
base_class=base_class,
sub_class=plugin.plugin,
reason=f"Entry point reference not a subclass of '{base_class}'.",
)
continue
_process_subclass(
sub_class=cls,
base_class=base_class,
type_id_key=type_id_key,
type_id_func=type_id_func,
type_id_no_attach=type_id_no_attach,
attach_python_metadata=attach_python_metadata,
ignore_abstract_classes=ignore_abstract_classes,
)
result_entrypoints[name] = cls
elif (
isinstance(plugin.plugin, tuple)
and len(plugin.plugin) >= 1
and callable(plugin.plugin[0])
) or callable(plugin.plugin):
try:
if callable(plugin.plugin):
func = plugin.plugin
args = []
else:
func = plugin.plugin[0]
args = plugin.plugin[1:]
classes = func(*args)
except Exception as e:
if is_debug():
import traceback
traceback.print_exc()
raise Exception(f"Error trying to load plugin '{plugin.plugin}': {e}")
for sub_class in classes:
type_id = _process_subclass(
sub_class=sub_class,
base_class=base_class,
type_id_key=type_id_key,
type_id_func=type_id_func,
type_id_no_attach=type_id_no_attach,
attach_python_metadata=attach_python_metadata,
ignore_abstract_classes=ignore_abstract_classes,
)
if type_id is None:
continue
if type_id in result_dynamic.keys():
raise Exception(
f"Duplicate item name for type {entry_point_name}: {type_id}"
)
result_dynamic[type_id] = sub_class
else:
raise Exception(
f"Can't load subclasses for entry point {entry_point_name} and base class {base_class}: invalid plugin type {type(plugin.plugin)}"
)
for k, v in result_dynamic.items():
if k in result_entrypoints.keys():
msg = f"Duplicate item name '{k}' for type {entry_point_name}: {v} -- {result_entrypoints[k]}."
try:
if type_id_key not in v.__dict__.keys():
msg = f"{msg} Most likely the name is picked up from a subclass, try to add a '{type_id_key}' class attribute to your implementing class, with the name you want to give your type as value."
except Exception:
pass
raise Exception(msg)
result_entrypoints[k] = v
return result_entrypoints
cli
¶
F
¶FC
¶Classes¶
OutputFormat (Enum)
¶An enumeration.
Source code in kiara/utils/cli.py
class OutputFormat(Enum):
@classmethod
def as_dict(cls):
return {i.name: i.value for i in cls}
@classmethod
def keys_as_list(cls):
return cls._member_names_
@classmethod
def values_as_list(cls):
return [i.value for i in cls]
TERMINAL = "terminal"
HTML = "html"
JSON = "json"
JSON_INCL_SCHEMA = "json-incl-schema"
JSON_SCHEMA = "json-schema"
Functions¶
output_format_option(*param_decls)
¶Attaches an option to the command. All positional arguments are
passed as parameter declarations to :class:Option; all keyword
arguments are forwarded unchanged (except cls).
This is equivalent to creating an :class:Option instance manually
and attaching it to the :attr:Command.params list.
:param cls: the option class to instantiate. This defaults to
:class:Option.
Source code in kiara/utils/cli.py
def output_format_option(*param_decls: str) -> Callable[[FC], FC]:
"""Attaches an option to the command. All positional arguments are
passed as parameter declarations to :class:`Option`; all keyword
arguments are forwarded unchanged (except ``cls``).
This is equivalent to creating an :class:`Option` instance manually
and attaching it to the :attr:`Command.params` list.
:param cls: the option class to instantiate. This defaults to
:class:`Option`.
"""
if not param_decls:
param_decls = ("--format", "-f")
attrs = {
"help": "The output format. Defaults to 'terminal'.",
"type": click.Choice(OutputFormat.values_as_list()),
}
def decorator(f: FC) -> FC:
# Issue 926, copy attrs, so pre-defined options can re-use the same cls=
option_attrs = attrs.copy()
OptionClass = option_attrs.pop("cls", None) or Option
_param_memo(f, OptionClass(param_decls, **option_attrs)) # type: ignore
return f
return decorator
render_json_schema_str(model)
¶Source code in kiara/utils/cli.py
def render_json_schema_str(model: BaseModel):
try:
json_str = model.schema_json(option=orjson.OPT_INDENT_2)
except TypeError:
json_str = model.schema_json(indent=2)
return json_str
render_json_str(model)
¶Source code in kiara/utils/cli.py
def render_json_str(model: BaseModel):
try:
json_str = model.json(option=orjson.OPT_INDENT_2)
except TypeError:
json_str = model.json(indent=2)
return json_str
terminal_print(msg=None, in_panel=None, rich_config=None, empty_line_before=False, **config)
¶Source code in kiara/utils/cli.py
def terminal_print(
msg: Any = None,
in_panel: Optional[str] = None,
rich_config: Optional[Mapping[str, Any]] = None,
empty_line_before: bool = False,
**config: Any,
) -> None:
if msg is None:
msg = ""
console = get_console()
msg = extract_renderable(msg, render_config=config)
# if hasattr(msg, "create_renderable"):
# msg = msg.create_renderable(**config) # type: ignore
if in_panel is not None:
msg = Panel(msg, title_align="left", title=in_panel)
if empty_line_before:
console.print()
if rich_config:
console.print(msg, **rich_config)
else:
console.print(msg)
terminal_print_model(*models, *, format=None, empty_line_before=None, in_panel=None, **render_config)
¶Source code in kiara/utils/cli.py
def terminal_print_model(
*models: BaseModel,
format: Union[None, OutputFormat, str] = None,
empty_line_before: Optional[bool] = None,
in_panel: Optional[str] = None,
**render_config: Any,
):
if format is None:
format = OutputFormat.TERMINAL
if isinstance(format, str):
format = OutputFormat(format)
if empty_line_before is None:
if format == OutputFormat.TERMINAL:
empty_line_before = True
else:
empty_line_before = False
if format == OutputFormat.TERMINAL:
if len(models) == 1:
terminal_print(
models[0],
in_panel=in_panel,
empty_line_before=empty_line_before,
**render_config,
)
else:
rg = []
for model in models[0:-1]:
renderable = extract_renderable(model, render_config)
rg.append(renderable)
rg.append(Rule(style="b"))
last = extract_renderable(models[-1], render_config)
rg.append(last)
group = Group(*rg)
terminal_print(group, in_panel=in_panel, **render_config)
elif format == OutputFormat.JSON:
if len(models) == 1:
json_str = render_json_str(models[0])
syntax = Syntax(json_str, "json", background_color="default")
terminal_print(
syntax,
empty_line_before=empty_line_before,
rich_config={"soft_wrap": True},
)
else:
json_strs = []
for model in models:
json_str = render_json_str(model)
json_strs.append(json_str)
json_str_full = "[" + ",\n".join(json_strs) + "]"
syntax = Syntax(json_str_full, "json", background_color="default")
terminal_print(
syntax,
empty_line_before=empty_line_before,
rich_config={"soft_wrap": True},
)
elif format == OutputFormat.JSON_SCHEMA:
if len(models) == 1:
syntax = Syntax(
models[0].schema_json(option=orjson.OPT_INDENT_2),
"json",
background_color="default",
)
terminal_print(
syntax,
empty_line_before=empty_line_before,
rich_config={"soft_wrap": True},
)
else:
json_strs = []
for model in models:
json_strs.append(render_json_schema_str(model))
json_str_full = "[" + ",\n".join(json_strs) + "]"
syntax = Syntax(json_str_full, "json", background_color="default")
terminal_print(
syntax,
empty_line_before=empty_line_before,
rich_config={"soft_wrap": True},
)
elif format == OutputFormat.JSON_INCL_SCHEMA:
if len(models) == 1:
data = models[0].dict()
schema = models[0].schema()
all = {"data": data, "schema": schema}
json_str = orjson_dumps(all, option=orjson.OPT_INDENT_2)
syntax = Syntax(json_str, "json", background_color="default")
terminal_print(
syntax,
empty_line_before=empty_line_before,
rich_config={"soft_wrap": True},
)
else:
all_data = []
for model in models:
data = model.dict()
schema = model.schema()
all_data.append({"data": data, "schema": schema})
json_str = orjson_dumps(all_data, option=orjson.OPT_INDENT_2)
# print(json_str)
syntax = Syntax(json_str, "json", background_color="default")
terminal_print(
syntax,
empty_line_before=empty_line_before,
rich_config={"soft_wrap": True},
)
elif format == OutputFormat.HTML:
all_html = ""
for model in models:
if hasattr(model, "create_html"):
html = model.create_html() # type: ignore
all_html = f"{all_html}\n{html}"
else:
raise NotImplementedError()
syntax = Syntax(all_html, "html", background_color="default")
terminal_print(
syntax, empty_line_before=empty_line_before, rich_config={"soft_wrap": True}
)
concurrency
¶
Classes¶
ThreadSaveCounter
¶A thread-safe counter, can be used in kiara modules to update completion percentage.
Source code in kiara/utils/concurrency.py
class ThreadSaveCounter(object):
"""A thread-safe counter, can be used in kiara modules to update completion percentage."""
def __init__(self):
self._current = 0
self._lock = threading.Lock()
@property
def current(self):
return self._current
def current_percent(self, total: int) -> int:
return int((self.current / total) * 100)
def increment(self):
with self._lock:
self._current += 1
return self._current
def decrement(self):
with self._lock:
self._current -= 1
return self._current
current
property
readonly
¶current_percent(self, total)
¶Source code in kiara/utils/concurrency.py
def current_percent(self, total: int) -> int:
return int((self.current / total) * 100)
decrement(self)
¶Source code in kiara/utils/concurrency.py
def decrement(self):
with self._lock:
self._current -= 1
return self._current
increment(self)
¶Source code in kiara/utils/concurrency.py
def increment(self):
with self._lock:
self._current += 1
return self._current
data
¶
logger
¶render_data(kiara, value_id, target_type='terminal_renderable', **render_config)
¶Source code in kiara/utils/data.py
def render_data(
kiara: "Kiara",
value_id: uuid.UUID,
target_type="terminal_renderable",
**render_config: Any,
) -> Any:
value = kiara.data_registry.get_value(value_id=value_id)
op_type: RenderValueOperationType = kiara.operation_registry.get_operation_type("render_value") # type: ignore
try:
op: Optional[Operation] = op_type.get_operation_for_render_combination(
source_type=value.value_schema.type, target_type=target_type
)
except Exception as e:
logger.debug(
"error.render_value",
source_type=value.value_schema.type,
target_type=target_type,
error=e,
)
op = None
if target_type == "terminal_renderable":
try:
op = op_type.get_operation_for_render_combination(
source_type="any", target_type="string"
)
except Exception:
pass
if op is None:
raise Exception(
f"Can't find operation to render '{value.value_schema.type}' as '{target_type}."
)
assert op is not None
result = op.run(kiara=kiara, inputs={"value": value})
rendered = result.get_value_data("rendered_value")
return rendered
db
¶
get_kiara_db_url(base_path)
¶Source code in kiara/utils/db.py
def get_kiara_db_url(base_path: str):
abs_path = os.path.abspath(os.path.expanduser(base_path))
db_url = f"sqlite+pysqlite:///{abs_path}/kiara.db"
return db_url
orm_json_deserialize(obj)
¶Source code in kiara/utils/db.py
def orm_json_deserialize(obj: str) -> Any:
return orjson.loads(obj)
orm_json_serialize(obj)
¶Source code in kiara/utils/db.py
def orm_json_serialize(obj: Any) -> str:
if hasattr(obj, "json"):
return obj.json()
if isinstance(obj, str):
return obj
elif isinstance(obj, Mapping):
return orjson_dumps(obj, default=None)
else:
raise Exception(f"Unsupported type for json serialization: {type(obj)}")
doc
¶
extract_doc_from_cls(cls, only_first_line=False)
¶Source code in kiara/utils/doc.py
def extract_doc_from_cls(cls: typing.Type, only_first_line: bool = False):
doc = cls.__doc__
if not doc:
doc = DEFAULT_NO_DESC_VALUE
else:
doc = cleandoc(doc)
if only_first_line:
return first_line(doc)
else:
return doc.strip()
global_metadata
¶
get_metadata_for_python_module_or_class(module_or_class)
¶Source code in kiara/utils/global_metadata.py
@lru_cache()
def get_metadata_for_python_module_or_class(
module_or_class: typing.Union[ModuleType, typing.Type]
) -> typing.List[typing.Dict[str, typing.Any]]:
metadata: typing.List[typing.Dict[str, typing.Any]] = []
if isinstance(module_or_class, type):
if hasattr(module_or_class, KIARA_MODULE_METADATA_ATTRIBUTE):
md = getattr(module_or_class, KIARA_MODULE_METADATA_ATTRIBUTE)
assert isinstance(md, typing.Mapping)
metadata.append(md) # type: ignore
_module_or_class: typing.Union[
str, ModuleType, typing.Type
] = module_or_class.__module__
else:
_module_or_class = module_or_class
current_module = _module_or_class
while current_module:
if isinstance(current_module, str):
current_module = importlib.import_module(current_module)
if hasattr(current_module, KIARA_MODULE_METADATA_ATTRIBUTE):
md = getattr(current_module, KIARA_MODULE_METADATA_ATTRIBUTE)
assert isinstance(md, typing.Mapping)
metadata.append(md) # type: ignore
if "." in current_module.__name__:
current_module = ".".join(current_module.__name__.split(".")[0:-1])
else:
current_module = ""
metadata.reverse()
return metadata
graphs
¶
print_ascii_graph(graph)
¶Source code in kiara/utils/graphs.py
def print_ascii_graph(graph: nx.Graph):
try:
from asciinet import graph_to_ascii
except: # noqa
print(
"\nCan't print graph on terminal, package 'asciinet' not available. Please install it into the current virtualenv using:\n\npip install 'git+https://github.com/cosminbasca/asciinet.git#egg=asciinet&subdirectory=pyasciinet'"
)
return
try:
from asciinet._libutil import check_java
check_java("Java ")
except Exception as e:
print(e)
print(
"\nJava is currently necessary to print ascii graph. This might change in the future, but to use this functionality please install a JRE."
)
return
print(graph_to_ascii(graph))
hashing
¶
compute_hash(obj)
¶Source code in kiara/utils/hashing.py
def compute_hash(obj: Any) -> int:
h = DeepHash(obj, hasher=KIARA_HASH_FUNCTION)
return h[obj]
jupyter
¶
create_image(graph)
¶Source code in kiara/utils/jupyter.py
def create_image(graph: nx.Graph):
try:
import pygraphviz as pgv # noqa
except: # noqa
return "pygraphviz not available, please install it manually into the current virtualenv"
# graph = nx.relabel_nodes(graph, lambda x: hash(x))
G = nx.nx_agraph.to_agraph(graph)
G.node_attr["shape"] = "box"
# G.unflatten().layout(prog="dot")
G.layout(prog="dot")
b = G.draw(format="png")
return b
graph_to_image(graph, return_bytes=False)
¶Source code in kiara/utils/jupyter.py
def graph_to_image(graph: nx.Graph, return_bytes: bool = False):
b = create_image(graph=graph)
if return_bytes:
return b
else:
return Image(b)
save_image(graph, path)
¶Source code in kiara/utils/jupyter.py
def save_image(graph: nx.Graph, path: str):
graph_b = create_image(graph=graph)
with open(path, "wb") as f:
f.write(graph_b)
metadata
¶
find_metadata_models(alias=None, only_for_package=None)
¶Source code in kiara/utils/metadata.py
def find_metadata_models(
alias: Optional[str] = None, only_for_package: Optional[str] = None
) -> MetadataTypeClassesInfo:
models = find_all_value_metadata_models()
group: MetadataTypeClassesInfo = MetadataTypeClassesInfo.create_from_type_items(group_alias=alias, **models) # type: ignore
if only_for_package:
temp = {}
for key, info in group.items():
if info.context.labels.get("package") == only_for_package:
temp[key] = info
group = MetadataTypeClassesInfo.construct(
group_id=group.group_id, group_alias=group.group_alias, type_infos=temp # type: ignore
)
return group
models
¶
Functions¶
assemble_subcomponent_tree(data)
¶Source code in kiara/utils/models.py
def assemble_subcomponent_tree(data: "KiaraModel") -> Optional[nx.DiGraph]:
graph = nx.DiGraph()
def assemble_tree(info_model: KiaraModel, current_node_id, level: int = 0):
graph.add_node(current_node_id, obj=info_model, level=level)
scn = info_model.subcomponent_keys
if not scn:
return
for child_path in scn:
child_obj = info_model.get_subcomponent(child_path)
new_node_id = f"{current_node_id}.{child_path}"
graph.add_edge(current_node_id, new_node_id)
assemble_tree(child_obj, new_node_id, level + 1)
assemble_tree(data, KIARA_DEFAULT_ROOT_NODE_ID)
return graph
create_pydantic_model(model_cls, _use_pydantic_construct=False, **field_values)
¶Source code in kiara/utils/models.py
def create_pydantic_model(
model_cls: Type[BaseModel],
_use_pydantic_construct: bool = PYDANTIC_USE_CONSTRUCT,
**field_values: Any,
):
if _use_pydantic_construct:
raise NotImplementedError()
return model_cls.construct(**field_values)
else:
return model_cls(**field_values)
get_subcomponent_from_model(data, path)
¶Return subcomponents of a model under a specified path.
Source code in kiara/utils/models.py
def get_subcomponent_from_model(data: "KiaraModel", path: str) -> "KiaraModel":
"""Return subcomponents of a model under a specified path."""
if "." in path:
first_token, rest = path.split(".", maxsplit=1)
sc = data.get_subcomponent(first_token)
return sc.get_subcomponent(rest)
if hasattr(data, "__custom_root_type__") and data.__custom_root_type__:
if isinstance(data.__root__, Mapping): # type: ignore
if path in data.__root__.keys(): # type: ignore
return data.__root__[path] # type: ignore
else:
matches = {}
for k in data.__root__.keys(): # type: ignore
if k.startswith(f"{path}."):
rest = k[len(path) + 1 :] # noqa
matches[rest] = data.__root__[k] # type: ignore
if not matches:
raise KeyError(f"No child models under '{path}'.")
else:
raise NotImplementedError()
# subcomponent_group = KiaraModelGroup.create_from_child_models(**matches)
# return subcomponent_group
else:
raise NotImplementedError()
else:
if path in data.__fields__.keys():
return getattr(data, path)
else:
raise KeyError(
f"No subcomponent for key '{path}' in model: {data.model_id}."
)
retrieve_data_subcomponent_keys(data)
¶Source code in kiara/utils/models.py
def retrieve_data_subcomponent_keys(data: Any) -> Iterable[str]:
if hasattr(data, "__custom_root_type__") and data.__custom_root_type__:
if isinstance(data.__root__, Mapping): # type: ignore
result = set()
for k, v in data.__root__.items(): # type: ignore
if isinstance(v, BaseModel):
result.add(k.split(".")[0])
return result
else:
return []
elif isinstance(data, BaseModel):
matches = []
for x in data.__fields__.keys():
_type = data.__fields__[x].type_
if isinstance(_type, type) and issubclass(_type, BaseModel):
matches.append(x)
return matches
else:
log_message(
f"No subcomponents retrieval supported for data of type: {type(data)}"
)
return []
operations
¶
filter_operations(kiara, pkg_name=None, **operations)
¶Source code in kiara/utils/operations.py
def filter_operations(
kiara: "Kiara", pkg_name: Optional[str] = None, **operations: "Operation"
) -> OperationGroupInfo:
result: Dict[str, OperationInfo] = {}
op_infos = kiara.operation_registry.get_context_metadata(only_for_package=pkg_name)
modules = kiara.module_registry.get_context_metadata(only_for_package=pkg_name)
for op_id, op in operations.items():
if op.module in modules:
result[op_id] = OperationInfo.create_from_operation(
kiara=kiara, operation=op
)
continue
opt_types = kiara.operation_registry.find_all_operation_types(op_id)
match = False
for ot in opt_types:
if ot in op_infos.keys():
match = True
break
if match:
result[op_id] = OperationInfo.create_from_operation(
kiara=kiara, operation=op
)
return OperationGroupInfo.construct(type_infos=result)
output
¶
Classes¶
ArrowTabularWrap (TabularWrap)
¶Source code in kiara/utils/output.py
class ArrowTabularWrap(TabularWrap):
def __init__(self, table: "ArrowTable"):
self._table: "ArrowTable" = table
super().__init__()
def retrieve_column_names(self) -> Iterable[str]:
return self._table.column_names
def retrieve_number_of_rows(self) -> int:
return self._table.num_rows
def slice(self, offset: int = 0, length: Optional[int] = None):
return self._table.slice(offset=offset, length=length)
def to_pydict(self) -> Mapping:
return self._table.to_pydict()
retrieve_column_names(self)
¶Source code in kiara/utils/output.py
def retrieve_column_names(self) -> Iterable[str]:
return self._table.column_names
retrieve_number_of_rows(self)
¶Source code in kiara/utils/output.py
def retrieve_number_of_rows(self) -> int:
return self._table.num_rows
slice(self, offset=0, length=None)
¶Source code in kiara/utils/output.py
def slice(self, offset: int = 0, length: Optional[int] = None):
return self._table.slice(offset=offset, length=length)
to_pydict(self)
¶Source code in kiara/utils/output.py
def to_pydict(self) -> Mapping:
return self._table.to_pydict()
DictTabularWrap (TabularWrap)
¶Source code in kiara/utils/output.py
class DictTabularWrap(TabularWrap):
def __init__(self, data: Mapping[str, Any]):
self._data: Mapping[str, Any] = data
def retrieve_number_of_rows(self) -> int:
return len(self._data)
def retrieve_column_names(self) -> Iterable[str]:
return self._data.keys()
def to_pydict(self) -> Mapping:
return self._data
def slice(self, offset: int = 0, length: Optional[int] = None) -> "TabularWrap":
result = {}
start = None
end = None
for cn in self._data.keys():
if start is None:
if offset > len(self._data):
return DictTabularWrap({cn: [] for cn in self._data.keys()})
start = offset
if not length:
end = len(self._data)
else:
end = start + length
if end > len(self._data):
end = len(self._data)
result[cn] = self._data[cn][start:end]
return DictTabularWrap(result)
retrieve_column_names(self)
¶Source code in kiara/utils/output.py
def retrieve_column_names(self) -> Iterable[str]:
return self._data.keys()
retrieve_number_of_rows(self)
¶Source code in kiara/utils/output.py
def retrieve_number_of_rows(self) -> int:
return len(self._data)
slice(self, offset=0, length=None)
¶Source code in kiara/utils/output.py
def slice(self, offset: int = 0, length: Optional[int] = None) -> "TabularWrap":
result = {}
start = None
end = None
for cn in self._data.keys():
if start is None:
if offset > len(self._data):
return DictTabularWrap({cn: [] for cn in self._data.keys()})
start = offset
if not length:
end = len(self._data)
else:
end = start + length
if end > len(self._data):
end = len(self._data)
result[cn] = self._data[cn][start:end]
return DictTabularWrap(result)
to_pydict(self)
¶Source code in kiara/utils/output.py
def to_pydict(self) -> Mapping:
return self._data
OutputDetails (BaseModel)
pydantic-model
¶Source code in kiara/utils/output.py
class OutputDetails(BaseModel):
@classmethod
def from_data(cls, data: Any):
if isinstance(data, str):
if "=" in data:
data = [data]
else:
data = [f"format={data}"]
if isinstance(data, Iterable):
data = list(data)
if len(data) == 1 and isinstance(data[0], str) and "=" not in data[0]:
data = [f"format={data[0]}"]
output_details_dict = dict_from_cli_args(*data)
else:
raise TypeError(
f"Can't parse output detail config: invalid input type '{type(data)}'."
)
output_details = OutputDetails(**output_details_dict)
return output_details
format: str = Field(description="The output format.")
target: str = Field(description="The output target.")
config: Dict[str, Any] = Field(
description="Output configuration.", default_factory=dict
)
@root_validator(pre=True)
def _set_defaults(cls, values):
target: str = values.pop("target", "terminal")
format: str = values.pop("format", None)
if format is None:
if target == "terminal":
format = "terminal"
else:
if target == "file":
format = "json"
else:
ext = target.split(".")[-1]
if ext in ["yaml", "json"]:
format = ext
else:
format = "json"
result = {"format": format, "target": target, "config": dict(values)}
return result
config: Dict[str, Any]
pydantic-field
¶Output configuration.
format: str
pydantic-field
required
¶The output format.
target: str
pydantic-field
required
¶The output target.
from_data(data)
classmethod
¶Source code in kiara/utils/output.py
@classmethod
def from_data(cls, data: Any):
if isinstance(data, str):
if "=" in data:
data = [data]
else:
data = [f"format={data}"]
if isinstance(data, Iterable):
data = list(data)
if len(data) == 1 and isinstance(data[0], str) and "=" not in data[0]:
data = [f"format={data[0]}"]
output_details_dict = dict_from_cli_args(*data)
else:
raise TypeError(
f"Can't parse output detail config: invalid input type '{type(data)}'."
)
output_details = OutputDetails(**output_details_dict)
return output_details
RenderConfig (BaseModel)
pydantic-model
¶
TabularWrap (ABC)
¶Source code in kiara/utils/output.py
class TabularWrap(ABC):
def __init__(self):
self._num_rows: Optional[int] = None
self._column_names: Optional[Iterable[str]] = None
@property
def num_rows(self) -> int:
if self._num_rows is None:
self._num_rows = self.retrieve_number_of_rows()
return self._num_rows
@property
def column_names(self) -> Iterable[str]:
if self._column_names is None:
self._column_names = self.retrieve_column_names()
return self._column_names
@abstractmethod
def retrieve_column_names(self) -> Iterable[str]:
pass
@abstractmethod
def retrieve_number_of_rows(self) -> int:
pass
@abstractmethod
def slice(self, offset: int = 0, length: Optional[int] = None) -> "TabularWrap":
pass
@abstractmethod
def to_pydict(self) -> Mapping:
pass
def pretty_print(
self,
rows_head: Optional[int] = None,
rows_tail: Optional[int] = None,
max_row_height: Optional[int] = None,
max_cell_length: Optional[int] = None,
) -> RenderableType:
rich_table = RichTable(box=box.SIMPLE)
for cn in self.retrieve_column_names():
rich_table.add_column(cn)
num_split_rows = 2
if rows_head is not None:
if rows_head < 0:
rows_head = 0
if rows_head > self.retrieve_number_of_rows():
rows_head = self.retrieve_number_of_rows()
rows_tail = None
num_split_rows = 0
if rows_tail is not None:
if rows_head + rows_tail >= self.num_rows: # type: ignore
rows_head = self.retrieve_number_of_rows()
rows_tail = None
num_split_rows = 0
else:
num_split_rows = 0
if rows_head is not None:
head = self.slice(0, rows_head)
num_rows = rows_head
else:
head = self
num_rows = self.retrieve_number_of_rows()
table_dict = head.to_pydict()
for i in range(0, num_rows):
row = []
for cn in self.retrieve_column_names():
cell = table_dict[cn][i]
cell_str = str(cell)
if max_row_height and max_row_height > 0 and "\n" in cell_str:
lines = cell_str.split("\n")
if len(lines) > max_row_height:
if max_row_height == 1:
lines = lines[0:1]
else:
half = int(max_row_height / 2)
lines = lines[0:half] + [".."] + lines[-half:]
cell_str = "\n".join(lines)
if max_cell_length and max_cell_length > 0:
lines = []
for line in cell_str.split("\n"):
if len(line) > max_cell_length:
line = line[0:max_cell_length] + " ..."
else:
line = line
lines.append(line)
cell_str = "\n".join(lines)
row.append(cell_str)
rich_table.add_row(*row)
if num_split_rows:
for i in range(0, num_split_rows):
row = []
for _ in self.retrieve_column_names():
row.append("...")
rich_table.add_row(*row)
if rows_head:
if rows_tail is not None:
if rows_tail < 0:
rows_tail = 0
tail = self.slice(self.retrieve_number_of_rows() - rows_tail)
table_dict = tail.to_pydict()
for i in range(0, num_rows):
row = []
for cn in self.retrieve_column_names():
cell = table_dict[cn][i]
cell_str = str(cell)
if max_row_height and max_row_height > 0 and "\n" in cell_str:
lines = cell_str.split("\n")
if len(lines) > max_row_height:
if max_row_height == 1:
lines = lines[0:1]
else:
half = int(len(lines) / 2)
lines = lines[0:half] + [".."] + lines[-half:]
cell_str = "\n".join(lines)
if max_cell_length and max_cell_length > 0:
lines = []
for line in cell_str.split("\n"):
if len(line) > max_cell_length:
line = line[0:(max_cell_length)] + " ..."
else:
line = line
lines.append(line)
cell_str = "\n".join(lines)
row.append(cell_str)
rich_table.add_row(*row)
return rich_table
column_names: Iterable[str]
property
readonly
¶num_rows: int
property
readonly
¶pretty_print(self, rows_head=None, rows_tail=None, max_row_height=None, max_cell_length=None)
¶Source code in kiara/utils/output.py
def pretty_print(
self,
rows_head: Optional[int] = None,
rows_tail: Optional[int] = None,
max_row_height: Optional[int] = None,
max_cell_length: Optional[int] = None,
) -> RenderableType:
rich_table = RichTable(box=box.SIMPLE)
for cn in self.retrieve_column_names():
rich_table.add_column(cn)
num_split_rows = 2
if rows_head is not None:
if rows_head < 0:
rows_head = 0
if rows_head > self.retrieve_number_of_rows():
rows_head = self.retrieve_number_of_rows()
rows_tail = None
num_split_rows = 0
if rows_tail is not None:
if rows_head + rows_tail >= self.num_rows: # type: ignore
rows_head = self.retrieve_number_of_rows()
rows_tail = None
num_split_rows = 0
else:
num_split_rows = 0
if rows_head is not None:
head = self.slice(0, rows_head)
num_rows = rows_head
else:
head = self
num_rows = self.retrieve_number_of_rows()
table_dict = head.to_pydict()
for i in range(0, num_rows):
row = []
for cn in self.retrieve_column_names():
cell = table_dict[cn][i]
cell_str = str(cell)
if max_row_height and max_row_height > 0 and "\n" in cell_str:
lines = cell_str.split("\n")
if len(lines) > max_row_height:
if max_row_height == 1:
lines = lines[0:1]
else:
half = int(max_row_height / 2)
lines = lines[0:half] + [".."] + lines[-half:]
cell_str = "\n".join(lines)
if max_cell_length and max_cell_length > 0:
lines = []
for line in cell_str.split("\n"):
if len(line) > max_cell_length:
line = line[0:max_cell_length] + " ..."
else:
line = line
lines.append(line)
cell_str = "\n".join(lines)
row.append(cell_str)
rich_table.add_row(*row)
if num_split_rows:
for i in range(0, num_split_rows):
row = []
for _ in self.retrieve_column_names():
row.append("...")
rich_table.add_row(*row)
if rows_head:
if rows_tail is not None:
if rows_tail < 0:
rows_tail = 0
tail = self.slice(self.retrieve_number_of_rows() - rows_tail)
table_dict = tail.to_pydict()
for i in range(0, num_rows):
row = []
for cn in self.retrieve_column_names():
cell = table_dict[cn][i]
cell_str = str(cell)
if max_row_height and max_row_height > 0 and "\n" in cell_str:
lines = cell_str.split("\n")
if len(lines) > max_row_height:
if max_row_height == 1:
lines = lines[0:1]
else:
half = int(len(lines) / 2)
lines = lines[0:half] + [".."] + lines[-half:]
cell_str = "\n".join(lines)
if max_cell_length and max_cell_length > 0:
lines = []
for line in cell_str.split("\n"):
if len(line) > max_cell_length:
line = line[0:(max_cell_length)] + " ..."
else:
line = line
lines.append(line)
cell_str = "\n".join(lines)
row.append(cell_str)
rich_table.add_row(*row)
return rich_table
retrieve_column_names(self)
¶Source code in kiara/utils/output.py
@abstractmethod
def retrieve_column_names(self) -> Iterable[str]:
pass
retrieve_number_of_rows(self)
¶Source code in kiara/utils/output.py
@abstractmethod
def retrieve_number_of_rows(self) -> int:
pass
slice(self, offset=0, length=None)
¶Source code in kiara/utils/output.py
@abstractmethod
def slice(self, offset: int = 0, length: Optional[int] = None) -> "TabularWrap":
pass
to_pydict(self)
¶Source code in kiara/utils/output.py
@abstractmethod
def to_pydict(self) -> Mapping:
pass
Functions¶
create_renderable_from_values(values, config=None)
¶Create a renderable for this module configuration.
Source code in kiara/utils/output.py
def create_renderable_from_values(
values: Mapping[str, "Value"], config: Optional[Mapping[str, Any]] = None
) -> RenderableType:
"""Create a renderable for this module configuration."""
if config is None:
config = {}
render_format = config.get("render_format", "terminal")
if render_format not in ["terminal"]:
raise Exception(f"Invalid render format: {render_format}")
show_pedigree = config.get("show_pedigree", False)
show_data = config.get("show_data", False)
show_hash = config.get("show_hash", True)
show_load_config = config.get("show_load_config", False)
table = RichTable(show_lines=True, box=box.MINIMAL_DOUBLE_HEAD)
table.add_column("value_id", "i")
table.add_column("data_type")
table.add_column("size")
if show_hash:
table.add_column("hash")
if show_pedigree:
table.add_column("pedigree")
if show_data:
table.add_column("data")
if show_load_config:
table.add_column("load_config")
for id, value in sorted(values.items(), key=lambda item: item[1].value_schema.type):
row: List[RenderableType] = [id, value.value_schema.type, str(value.value_size)]
if show_hash:
row.append(str(value.value_hash))
if show_pedigree:
if value.pedigree == ORPHAN:
pedigree = "-- n/a --"
else:
pedigree = value.pedigree.json(option=orjson.OPT_INDENT_2)
row.append(pedigree)
if show_data:
data = value._data_registry.render_data(
value_id=value.value_id, target_type="terminal_renderable", **config
)
row.append(data)
if show_load_config:
load_config = value.retrieve_load_config()
if load_config is None:
load_config_str: RenderableType = "-- not stored (yet) --"
else:
load_config_str = load_config.create_renderable()
row.append(load_config_str)
table.add_row(*row)
return table
create_table_from_base_model_cls(model_cls)
¶Source code in kiara/utils/output.py
def create_table_from_base_model_cls(model_cls: Type[BaseModel]):
table = RichTable(box=box.SIMPLE, show_lines=True)
table.add_column("Field")
table.add_column("Type")
table.add_column("Description")
table.add_column("Required")
table.add_column("Default")
props = model_cls.schema().get("properties", {})
for field_name, field in sorted(model_cls.__fields__.items()):
row = [field_name]
p = props.get(field_name, None)
p_type = None
if p is not None:
p_type = p.get("type", None)
# TODO: check 'anyOf' keys
if p_type is None:
p_type = "-- check source --"
row.append(p_type)
desc = p.get("description", "")
row.append(desc)
row.append("yes" if field.required else "no")
default = field.default
if callable(default):
default = default()
if default is None:
default = ""
else:
try:
default = json.dumps(default, indent=2)
except Exception:
default = str(default)
row.append(default)
table.add_row(*row)
return table
create_table_from_field_schemas(_add_default=True, _add_required=True, _show_header=False, _constants=None, **fields)
¶Source code in kiara/utils/output.py
def create_table_from_field_schemas(
_add_default: bool = True,
_add_required: bool = True,
_show_header: bool = False,
_constants: Optional[Mapping[str, Any]] = None,
**fields: "ValueSchema",
) -> RichTable:
table = RichTable(box=box.SIMPLE, show_header=_show_header)
table.add_column("field name", style="i")
table.add_column("type")
table.add_column("description")
if _add_required:
table.add_column("Required")
if _add_default:
if _constants:
table.add_column("Default / Constant")
else:
table.add_column("Default")
for field_name, schema in fields.items():
row: List[RenderableType] = [field_name, schema.type, schema.doc]
if _add_required:
req = schema.is_required()
if not req:
req_str = "no"
else:
if schema.default in [
None,
SpecialValue.NO_VALUE,
SpecialValue.NOT_SET,
]:
req_str = "[b]yes[b]"
else:
req_str = "no"
row.append(req_str)
if _add_default:
if _constants and field_name in _constants.keys():
d = f"[b]{_constants[field_name]}[/b] (constant)"
else:
if schema.default in [
None,
SpecialValue.NO_VALUE,
SpecialValue.NOT_SET,
]:
d = "-- no default --"
else:
d = str(schema.default)
row.append(d)
table.add_row(*row)
return table
create_table_from_model_object(model, render_config=None, exclude_fields=None)
¶Source code in kiara/utils/output.py
def create_table_from_model_object(
model: BaseModel,
render_config: Optional[Mapping[str, Any]] = None,
exclude_fields: Optional[Set[str]] = None,
):
model_cls = model.__class__
table = RichTable(box=box.SIMPLE, show_lines=True)
table.add_column("Field")
table.add_column("Type")
table.add_column("Value")
table.add_column("Description")
props = model_cls.schema().get("properties", {})
for field_name, field in sorted(model_cls.__fields__.items()):
if exclude_fields and field_name in exclude_fields:
continue
row = [field_name]
p = props.get(field_name, None)
p_type = None
if p is not None:
p_type = p.get("type", None)
# TODO: check 'anyOf' keys
if p_type is None:
p_type = "-- check source --"
row.append(p_type)
data = getattr(model, field_name)
row.append(extract_renderable(data, render_config=render_config))
desc = p.get("description", "")
row.append(desc)
table.add_row(*row)
return table
create_value_map_status_renderable(inputs, render_config=None)
¶Source code in kiara/utils/output.py
def create_value_map_status_renderable(
inputs: ValueMap, render_config: Optional[Mapping[str, Any]] = None
) -> RichTable:
if render_config is None:
render_config = {}
show_required: bool = render_config.get("show_required", True)
table = RichTable(box=box.SIMPLE, show_header=True)
table.add_column("field name", style="i")
table.add_column("status", style="b")
table.add_column("type")
table.add_column("description")
if show_required:
table.add_column("required")
invalid = inputs.check_invalid()
for field_name, value in inputs.items():
row: List[RenderableType] = [field_name]
if field_name in invalid.keys():
row.append(f"[red]{invalid[field_name]}[/red]")
else:
row.append("[green]valid[/green]")
row.extend([value.value_schema.type, value.value_schema.doc.description])
if show_required:
req = value.value_schema.is_required()
if not req:
req_str = "no"
else:
if value.value_schema.default in [
None,
SpecialValue.NO_VALUE,
SpecialValue.NOT_SET,
]:
req_str = "[b]yes[b]"
else:
req_str = "no"
row.append(req_str)
table.add_row(*row)
return table
extract_renderable(item, render_config=None)
¶Try to automatically find and extract or create an object that is renderable by the 'rich' library.
Source code in kiara/utils/output.py
def extract_renderable(item: Any, render_config: Optional[Mapping[str, Any]] = None):
"""Try to automatically find and extract or create an object that is renderable by the 'rich' library."""
if render_config is None:
render_config = {}
else:
render_config = dict(render_config)
inline_models_as_json = render_config.setdefault("inline_models_as_json", True)
if hasattr(item, "create_renderable"):
return item.create_renderable(**render_config)
elif isinstance(item, (ConsoleRenderable, RichCast, str)):
return item
elif isinstance(item, BaseModel) and not inline_models_as_json:
return create_table_from_model_object(item)
elif isinstance(item, BaseModel):
return item.json(indent=2)
elif isinstance(item, Mapping) and not inline_models_as_json:
table = RichTable(show_header=False, box=box.SIMPLE)
table.add_column("Key", style="i")
table.add_column("Value")
for k, v in item.items():
table.add_row(k, extract_renderable(v, render_config=render_config))
return table
elif isinstance(item, Mapping):
result = {}
for k, v in item.items():
if isinstance(v, BaseModel):
v = v.dict()
result[k] = v
return orjson_dumps(
result, option=orjson.OPT_INDENT_2 | orjson.OPT_NON_STR_KEYS
)
elif isinstance(item, Iterable):
_all = []
for i in item:
_all.append(extract_renderable(i))
rg = Group(*_all)
return rg
else:
return str(item)
pipelines
¶
Functions¶
check_doc_sidecar(path, data)
¶Source code in kiara/utils/pipelines.py
def check_doc_sidecar(
path: Union[Path, str], data: Mapping[str, Any]
) -> Mapping[str, Any]:
if isinstance(path, str):
path = Path(os.path.expanduser(path))
_doc = data["data"].get("documentation", None)
if _doc is None:
_doc_path = Path(path.as_posix() + ".md")
if _doc_path.is_file():
doc = _doc_path.read_text()
if doc:
data["data"]["documentation"] = doc
return data
create_step_value_address(value_address_config, default_field_name)
¶Source code in kiara/utils/pipelines.py
def create_step_value_address(
value_address_config: Union[str, Mapping[str, Any]],
default_field_name: str,
) -> "StepValueAddress":
if isinstance(value_address_config, StepValueAddress):
return value_address_config
sub_value: Optional[Mapping[str, Any]] = None
if isinstance(value_address_config, str):
tokens = value_address_config.split(".")
if len(tokens) == 1:
step_id = value_address_config
output_name = default_field_name
elif len(tokens) == 2:
step_id = tokens[0]
output_name = tokens[1]
elif len(tokens) == 3:
step_id = tokens[0]
output_name = tokens[1]
sub_value = {"config": tokens[2]}
else:
raise NotImplementedError()
elif isinstance(value_address_config, Mapping):
step_id = value_address_config["step_id"]
output_name = value_address_config["value_name"]
sub_value = value_address_config.get("sub_value", None)
else:
raise TypeError(
f"Invalid type for creating step value address: {type(value_address_config)}"
)
if sub_value is not None and not isinstance(sub_value, Mapping):
raise ValueError(
f"Invalid type '{type(sub_value)}' for sub_value (step_id: {step_id}, value name: {output_name}): {sub_value}"
)
input_link = StepValueAddress(
step_id=step_id, value_name=output_name, sub_value=sub_value
)
return input_link
ensure_step_value_addresses(link, default_field_name)
¶Source code in kiara/utils/pipelines.py
def ensure_step_value_addresses(
link: Union[str, Mapping, Iterable], default_field_name: str
) -> List["StepValueAddress"]:
if isinstance(link, (str, Mapping)):
input_links: List[StepValueAddress] = [
create_step_value_address(
value_address_config=link, default_field_name=default_field_name
)
]
elif isinstance(link, Iterable):
input_links = []
for o in link:
il = create_step_value_address(
value_address_config=o, default_field_name=default_field_name
)
input_links.append(il)
else:
raise TypeError(f"Can't parse input map, invalid type for output: {link}")
return input_links
get_pipeline_details_from_path(path, module_type_name=None, base_module=None)
¶Load a pipeline description, save it's content, and determine it the pipeline base name.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
path |
Union[str, pathlib.Path] |
the path to the pipeline file |
required |
module_type_name |
Optional[str] |
if specifies, overwrites any auto-detected or assigned pipeline name |
None |
base_module |
Optional[str] |
overrides the base module the assembled pipeline module will be located in the python hierarchy |
None |
Source code in kiara/utils/pipelines.py
def get_pipeline_details_from_path(
path: Union[str, Path],
module_type_name: Optional[str] = None,
base_module: Optional[str] = None,
) -> Mapping[str, Any]:
"""Load a pipeline description, save it's content, and determine it the pipeline base name.
Arguments:
path: the path to the pipeline file
module_type_name: if specifies, overwrites any auto-detected or assigned pipeline name
base_module: overrides the base module the assembled pipeline module will be located in the python hierarchy
"""
if isinstance(path, str):
path = Path(os.path.expanduser(path))
if not path.is_file():
raise Exception(
f"Can't add pipeline description '{path.as_posix()}': not a file"
)
data = get_data_from_file(path)
if not data:
raise Exception(
f"Can't register pipeline file '{path.as_posix()}': no content."
)
if module_type_name:
data[MODULE_TYPE_NAME_KEY] = module_type_name
if not isinstance(data, Mapping):
raise Exception("Not a dictionary type.")
# filename = path.name
# name = data.get(MODULE_TYPE_NAME_KEY, None)
# if name is None:
# name = filename.split(".", maxsplit=1)[0]
result = {"data": data, "source": path.as_posix(), "source_type": "file"}
if base_module:
result["base_module"] = base_module
return result
values
¶
augment_values(values, schemas, constants=None)
¶Source code in kiara/utils/values.py
def augment_values(
values: Mapping[str, Any],
schemas: Mapping[str, ValueSchema],
constants: Optional[Mapping[str, Any]] = None,
) -> Dict[str, Any]:
# TODO: check if extra fields were provided
if constants:
for k, v in constants.items():
if k in values.keys():
raise Exception(f"Invalid input: value provided for constant '{k}'")
values_new = {}
for field_name, schema in schemas.items():
if field_name not in values.keys():
if constants and field_name in constants:
values_new[field_name] = copy.deepcopy(constants[field_name])
elif schema.default != SpecialValue.NOT_SET:
if callable(schema.default):
values_new[field_name] = schema.default()
else:
values_new[field_name] = copy.deepcopy(schema.default)
else:
values_new[field_name] = SpecialValue.NOT_SET
else:
value = values[field_name]
if value is None:
value = SpecialValue.NO_VALUE
values_new[field_name] = value
return values_new
create_schema_dict(schema_config)
¶Source code in kiara/utils/values.py
def create_schema_dict(
schema_config: Mapping[str, Union[ValueSchema, Mapping[str, Any]]],
) -> Mapping[str, ValueSchema]:
invalid = check_valid_field_names(*schema_config.keys())
if invalid:
raise Exception(
f"Can't assemble schema because it contains invalid input field name(s) '{', '.join(invalid)}'. Change the input schema to not contain any of the reserved keywords: {', '.join(INVALID_VALUE_NAMES)}"
)
result = {}
for k, v in schema_config.items():
if isinstance(v, ValueSchema):
result[k] = v
elif isinstance(v, Mapping):
schema = ValueSchema(**v)
result[k] = schema
else:
if v is None:
msg = "None"
else:
msg = v.__class__
raise Exception(
f"Invalid return type '{msg}' for field '{k}' when trying to create schema."
)
return result
overlay_constants_and_defaults(schemas, defaults, constants)
¶Source code in kiara/utils/values.py
def overlay_constants_and_defaults(
schemas: Mapping[str, ValueSchema],
defaults: Mapping[str, Any],
constants: Mapping[str, Any],
):
for k, v in schemas.items():
default_value = defaults.get(k, None)
constant_value = constants.get(k, None)
# value_to_test = None
if default_value is not None and constant_value is not None:
raise Exception(
f"Module configuration error. Value '{k}' set in both 'constants' and 'defaults', this is not allowed."
)
# TODO: perform validation for constants/defaults
if default_value is not None:
schemas[k].default = default_value
if constant_value is not None:
schemas[k].default = constant_value
schemas[k].is_constant = True
input_schemas = {}
constants = {}
for k, v in schemas.items():
if v.is_constant:
constants[k] = v
else:
input_schemas[k] = v
return input_schemas, constants